Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.stsci.edu/~kgordon/papers/PS_files/ERE_DISM.ps.gz Äàòà èçìåíåíèÿ: Thu Apr 24 02:49:35 2003 Äàòà èíäåêñèðîâàíèÿ: Tue May 27 07:54:16 2008 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: earth's atmosphere

Detection of Extended Red Emisson in the Diffuse Interstellar Medium
Karl D. Gordon 1 , Adolf N. Witt, and Brian C. Friedmann
Ritter Astrophysical Research Center, The University of Toledo
Toledo, OH 43606
ABSTRACT
Extended Red Emission (ERE) has been detected in many dusty astrophysical
objects and this raises the question: Is ERE present only in discrete objects or is it
an observational feature of all dust, i.e. present in the diffuse interstellar medium?
In order to answer this question, we determined the blue and red intensities of the
radiation from the diffuse interstellar medium (ISM) and examined the red intensity
for the presence of an excess above that expected for scattered light. The diffuse ISM
blue and red intensities were obtained by subtracting the integrated star and galaxy
intensities from the blue and red measurements made by the Imaging Photopolarimeter
(IPP) aboard the Pioneer 10 and 11 spacecraft. The unique characteristic of the
Pioneer measurements is that they were taken outside the zodiacal dust cloud and,
therefore, are free from zodiacal light. The color of the diffuse ISM was found to be
redder than the Pioneer intensities. If the diffuse ISM intensities were entirely due to
scattering from dust (i.e. Diffuse Galactic Light or DGL), the color of the diffuse ISM
would be bluer than the Pioneer intensities. Finding a redder color implies the presence
of an excess red intensity. Using a model for the DGL, we found the blue diffuse
ISM intensity to be entirely attributable to the DGL. The red DGL was calculated
using the blue diffuse ISM intensities and the approximately invariant color of the
DGL calculated with the DGL model. Subtracting the calculated red DGL from the
red diffuse ISM intensities resulted in the detection of an excess red intensity with an
average value of ¸10 S 10 (V) G2V . This represents the likely detection of ERE in the
diffuse ISM since Hff emission cannot account for the strength of this excess and the
only other known emission process applicable to the diffuse ISM is ERE. Thus, ERE
appears to be a general characteristic of dust. The correlation between NHI and ERE
intensity is (1:43 \Sigma 0:31)\Theta10 \Gamma29 ergs s \Gamma1 š A \Gamma1 sr \Gamma1 H atom \Gamma1 from which the ERE
photon conversion efficiency was estimated at 10 \Sigma 3%.
1. Introduction
Extended Red Emission (ERE) is a broad (\Delta– ¸ 800 š A) emission band with a peak
wavelength between 6500 š A and 8000 š A seen in many dusty astrophysical objects (Figure 1). For
1 present address: Department of Physics & Astronomy, Louisiana State University, Baton Rouge, LA 70803

-- 2 --
many years, the Red Rectangle was the only object known to possess such an emission feature (see
Schmidt, Cohen, & Margon [1980] for an excellent Red Rectangle spectrum). This all changed
with the discovery that many reflection nebulae possessed flux levels in the R and I bands in
excess of that expected from dust scattered starlight (Witt, Schild, & Kraiman 1984; Witt &
Schild 1985, 1986). The spectroscopic confirmation of the excess flux as ERE (Witt & Schild 1988;
Witt & Malin 1989; Witt et al. 1989; Witt & Boroson 1990) proved that ERE was a feature of
many, but not all, reflection nebulae. The identification of the band as emission was strengthened
by the imaging polarimetry of NGC 7023 by Watkin, Geldhill, & Scarrott (1991), which showed a
reduction in the R and I polarization where R and I excess flux existed. This convincingly proved
that the excess flux is due to an emission feature and not changes in the dust scattering properties.
Fig. 1.--- An example of Extended Red Emission is plotted. This spectrum is for the reflection
nebula NGC 2327 and has had the underlying scattering continuum subtracted (Witt 1988).
The detection of ERE in other dusty astrophysical objects quickly followed the confirmation
of ERE in reflection nebulae. Up to the date of this paper, ERE has been detected in the Red
Rectangle (Schmidt et al. 1980), reflection nebulae (see above), a dark nebula (Mattila 1979;

-- 3 --
Chlewicki & Laureijs 1987), Galactic cirrus clouds (Guhathakurta & Tyson 1989; Guhathakurta
& Cutri 1994; Szomoru & Guhathakurta 1998), planetary nebulae (Furton & Witt 1990, 1992),
H II regions (Perrin & Sivan 1992; Sivan & Perrin 1993), a nova (Scott, Evans, & Rawlings 1994),
the halo of the galaxy M82 (Perrin, Darbon, & Sivan 1995), and the 30 Doradus nebula in the
Large Magellanic Cloud (Darbon, Perrin, & Sivan 1998).
Clues as to the identity of the material that produces ERE are contained in the above
observations. The first clue comes from the wide variety of objects that show ERE. The material
must be able to survive in radically different environments: from cold, quiescent environments
(dark nebulae and Galactic cirrus clouds) to hot, dynamic environments (reflection nebulae,
planetary nebulae, and H II regions). Second, the material most likely is carbonaceous since ERE
has been detected in carbon rich planetary nebulae but not oxygen rich planetary nebulae (Furton
& Witt 1992). The robustness of carbonaceous dust material is supported by the essentially
constant C/H values found for sight lines exhibiting different dust characteristics (Sofia et al.
1997), implying that carbon is not easily exchanged between the gas and dust phases of the ISM.
The third clue comes from the spatial distribution of ERE in reflection nebulae (Witt & Schild
1988; Witt & Malin 1989; Witt et al. 1989; Witt & Boroson 1990; Rogers, Heyer, & Dewdney
1995; Lemaire et al. 1996) and planetary nebulae (Furton & Witt 1990). The ERE is strongest
in regions where H 2 is being dissociated, leading to the conclusion that warm atomic hydrogen
increases the efficiency of ERE luminescence. So, the material that produces ERE must be robust,
contain carbon, and produce ERE more efficiently in the presence of atomic hydrogen.
A prime candidate for this material is hydrogenated amorphous carbon (HAC), which was
first proposed to explain the ERE in the Red Rectangle (Duley 1985). With the discovery of ERE
in reflection nebulae and other dusty objects, the identification of HAC with ERE has strengthened
(Duley & Williams 1988; Witt & Schild 1988; Duley & Williams 1990; Witt & Furton 1994). The
identification of HAC with ERE is strongly supported by the laboratory work of Furton & Witt
(1993). They have shown that the low or non­existent photoluminescence of previously annealed
HAC or pure amorphous carbon can be greatly enhanced by exposure to atomic hydrogen and/or
ultraviolet radiation. This corresponds to the strengthening of ERE in H 2 dissociation regions
discussed above. Other ERE producing materials have been proposed, such as filmy quenched
carbonaceous composite (Sakata et al. 1992) and C 60 (Webster 1993). For a discussion of the
similarities and differences between these materials see Papoular et al. (1996).
As ERE has been detected in a large range of dusty objects, the question arises: Is ERE
present in the diffuse interstellar medium (ISM)? If the answer is yes, this would make ERE a
general characteristic of dust. A positive answer has been claimed by Duley & Whittet (1990),
who identified the very broad structure (VBS) seen in extinction curves (e.g. van Breda & Whittet
1981) as due to ERE. Jenniskens (1994) has pointed out that ERE cannot be the cause of the
VBS for two reasons. First, spectra showing VBS have had the nearby sky subtracted, effectively
removing any emission from the diffuse ISM. Second, the VBS strength does not depend on the
size of the aperture used. Hence, ERE has not been detected in the diffuse ISM. As a result, dust

-- 4 --
models for the diffuse ISM have not used the ability to produce ERE as a constraint on possible
dust grain materials (e.g. Kim & Martin 1996; Mathis 1996; Zubko, Kre/lowski, Wegner 1996;
Dwek et al. 1997; Li & Greenberg 1997).
This investigation is aimed at determining whether ERE is present in the diffuse ISM.
Detecting ERE in the diffuse ISM is much more difficult than doing the same in a discrete
object. For a discrete object, ERE detection is done by subtracting a nearby sky spectrum from a
spectrum of the object (e.g. Witt & Boroson 1990). Subtracting a nearby sky spectrum removes
contributions to the object spectrum from the Earth's atmosphere (airglow), zodiacal light (dust
scattered sunlight), and Galactic background light (diffuse ISM, faint stars, and galaxies). This
results in a spectrum with contributions only from the object being studied, and any ERE is
directly attributable to that object. As the light from the diffuse ISM is part of the sky spectrum,
this method will not work for it. A different method is required.
Two of the strongest (and most difficult to model) sources in a sky spectrum are airglow and
zodiacal light (Toller 1981). Both of these sources can be avoided by simply taking observations
outside the atmosphere (for airglow) and the zodiacal dust cloud (for zodiacal light). Such
measurements have already been carried out by the Imaging Photopolarimeters (IPP) carried
aboard both Pioneer 10 and 11 (Pellicori et al. 1973; Weinberg et al. 1974). The IPP measured
the intensity of almost the entire sky in the blue (437 nm) and the red (644 nm). By using only
measurements taken when the Pioneer spacecraft were beyond 3.27 AU, contributions from the
zodiacal light are avoided (Hanner et al. 1974). Therefore, the only known sources contributing to
the IPP measurements are stars, galaxies, and the diffuse ISM. Using photometric star and galaxy
catalogs, the contribution to the IPP measurements from stars and galaxies can be removed. The
resulting blue and red intensities are due only to the diffuse ISM.
While the IPP measurements have given all­sky maps at two wavelengths and not a spectrum,
this is sufficient to detect ERE in the diffuse ISM. The presence of ERE in reflection nebulae
was first detected by observing that these objects had red fluxes in excess of that expected from
dust scattered starlight (Witt, Schild, & Kraiman 1984). The diffuse ISM is a gigantic reflection
nebula with the Galaxy's starlight scattered by the Galaxy's dust. Therefore, we can use the same
criterion, excess red flux, to detect ERE in the diffuse ISM. The scattered light in the diffuse ISM
is termed Diffuse Galactic Light (DGL). The DGL will have a bluer color than the integrated
starlight because scattering by dust is more efficient at shorter wavelengths. So, if the diffuse ISM
color (red/blue ratio) is as red as or redder than the integrated starlight and other sources of
excess red light can be positively excluded, ERE is present.
Section 2 describes the Pioneer IPP measurements and the construction of the blue and
red all­sky maps. The compilation of a star and galaxy photometric catalog, complete to
approximately 20th magnitude, is detailed in section 3. The detection of ERE in the diffuse ISM
is contained in section 4. Section 5 presents the properties of the ERE in the diffuse ISM. Finally,
section 6 discusses the implications of our results and summarizes our conclusions.

-- 5 --
2. Pioneer Data
One of the instruments onboard the Pioneer 10 and 11 spacecraft was the Imaging
Photopolarimeter (IPP). The primary objectives of the IPP were to produce blue and red maps of
the brightness and polarization of the zodiacal dust cloud from 1 to 5 AU, the background light
outside the zodiacal dust cloud, and Jupiter (Pellicori et al. 1973). Of these, we were concerned
with only the all­sky blue and red surface brightness maps taken outside the zodiacal dust cloud.
The IPP was a Maksutov­type f/3.4 telescope with an aperture of 2.54 cm and a detector
consisting of a Wollaston prism, multilayer filters, and two dual­channel Bendix channeltrons
(Pellicori et al. 1973; Weinberg et al. 1974). Simultaneous measurements were made of the
orthogonal components of the electric field in both the blue and red. The spectral bandpass
(half­power) was 3950--4850 š A for the blue channel and 5900--6900 š A for the red channel (Pellicori
et al. 1973). See subsection 3.1 for more information on the photometric characteristics of
the IPP. The IPP instantaneous field of view (FOV) was 2: ffi 29 \Theta 2: ffi 29 for the background light
measurements. The IPP was mounted on a movable arm and 64 measurements were taken during
a single rotation of the Pioneer spacecraft. The angle between the arm and the Pioneer spacecraft
spin axis (look angle = L) was changed in increments of 1: ffi 83 to build up a map of the sky.
The look angle ranged between 29 ffi and 170 ffi (Pellicori et al. 1973). The effective FOV of the
measurements was 2: ffi 29 \Theta (2: ffi 29 + 5: ffi 625 sin L), with a maximum of 2: ffi 29 \Theta 7: ffi 92 when the look angle
was 90 ffi and a minimum of 2: ffi 29 \Theta 3: ffi 27 when the look angle was 170 ffi . At each look angle, a 20
data roll (rotation) measurement cycle was performed with 8 rolls for the background light, 1 for a
radioisotope­activated phosphor source ( 14 C), 1 for offset and dark current levels, and 10 for data
readout (Pellicori et al. 1973; Weinberg et al. 1974; Toller 1981).
The raw IPP background sky measurements were processed to produce the Pioneer 10/11
Background Sky data set available from the National Space Science Data Center (NSSDC).
The details of the processing can be found elsewhere (Weinberg et al. 1974; Toller 1981;
Weinberg & Schuerman 1981; Schuerman, Giovane, & Weinberg 1997). A brief description
of the processing follows. First, the data were calibrated using the inflight measurements of
the radioisotope­activated phosphor source. Second, the FOV center was computed from the
spacecraft spin axis direction, the look angle, and the clock angle. Third, the contribution from
bright stars was subtracted using the stars in the Bright Star Catalogue (Hoffleit & Warren 1991)
and stars with m V ! 8 (Toller, Tanabe, & Weinberg 1987) in the Photoelectric Catalog (Blanco
et al. 1968; Ochsenbein 1974). Fourth, 37 resolved stars were used to determine the time decay
of the instrument sensitivity and corrections to the telescope pointing. The final error in the
positions of the FOVs was on the order of 0: ffi 15--0: ffi 40. The Pioneer 10 red data have abnormally
high noise, but as we are also using Pioneer 11 data, this did not adversely affect our results. The
final Pioneer data are expressed in S 10 (V) G2V units, the equivalent number of 10th magnitude (V
band) solar­type stars per square degree. See subsection 3.1 for details of this unit.
During the cruise portion of the Pioneer 10 and 11 missions, the IPP mapped the background

-- 6 --
light a number of times. In order to determine the spatial extent of the zodiacal dust cloud,
Hanner et al. (1974) examined the brightness of two different regions of the sky as seen by the
IPP when Pioneer 10 was between 2.41 and 4.82 AU. They found that the brightness of these two
regions stopped changing after Pioneer 10 passed 3.27 AU, making this distance the outermost
detectable edge of the zodiacal dust cloud. Therefore, all the measurements taken beyond 3.27 AU
are free from detectable zodiacal light and useful for this investigation.
On 5 days while Pioneer 10 was between 3.26 and 5.15 AU and on 6 days while Pioneer 11
was between 4.06 and 4.66 AU, the IPP mapped the background light in the sky. The resolution of
a map made on a single day is determined by the FOV of the IPP and its overlap with neighboring
FOVs. Figure 2a gives an example of the pattern of FOVs using the Pioneer 10 measurements from
day 68 of 1974. One of the FOVs has been shaded to show the overlap of a FOV with neighboring
FOVs. The resolution of this map is variable, with each parallelogram being a resolution element.
In order to actually achieve this theoretical resolution, an algorithm must be used to extract the
information in the overlapping regions. In fact, the resolution of the IPP measurements can be
increased significantly by using measurements made on different days. Figure 2b gives the pattern
of FOVs for the 11 days used in creating the final high­resolution maps (see below). The FOVs
from different days do not overlap exactly as the Pioneer 10 and 11 missions were launched on
different trajectories and the spin axis of each spacecraft changed direction slowly as a function of
distance from the Sun.
Fig. 2.--- The pattern of the FOVs are plotted in these two figures. The pattern of FOVs from
Pioneer 10 on day 68 of 1974 is shown in (a) with points in the center of the FOVs. One FOV
is shaded to show the regions which overlap neighboring FOVs. The pattern of FOVs from the
11 days used in constructing the final maps is shown in (b). Due to the different trajectories of
Pioneer 10 and 11 and the variable spacecraft spin axis orientation, the FOVs from different days
do not overlap exactly.

-- 7 --
2.1. Map Generation Algorithm
The algorithm used to create the final maps is similar to that by Aumann, Fowler, & Melnyk
(1990). Their algorithm is called the Maximum Correlation Method (MCM) and it was able to
improve the resolution of IRAS maps by a factor of ¸ 6:5, from ¸ 4 0 to ¸ 36 00 . Our algorithm
was similar to MCM and worked in the following manner. An initial guess at the final image
(zeroth iteration) was taken as a positive flat image with 0: ffi 25 \Theta 0: ffi 25 pixels. The next iteration
was calculated from
p k+1
ij =
/
1
N
N
X
m=1
Cm
!
p k
ij (1)
where p k
ij is the surface brightness of the kth iteration image at pixel coordinates (i,j), Cm is the
correction factor for the mth IPP measurement (the surface brightness in particular FOV) which
includes p ij , and the sum was done over the N IPP measurements which include p ij . The value of
Cm was calculated from
Cm = I m
0
@ 1
Q
Q
X
p k
ij
1
A
\Gamma1
(2)
where I m is the mth IPP measurement and the sum was done over the Q pixels which are included
in the mth IPP FOV. The error, oe k
ij , in p k
ij was calculated from
oe k
ij = 1
N
v u u u t N
X
m=1
0
@ I m \Gamma
2
4 1
Q
Q
X
p k
ij
3
5
m
1
A
2
: (3)
With each iteration, the image gives an improved match to the IPP measurements. The outcome
of this algorithm is to produce an image which describes all 11 days of the Pioneer measurements.
The 11 days of IPP background light measurements that were used in constructing the final
high­resolution maps are tabulated in Table 1. While the 11 days overall possessed usable data,
a large number of individual measurements were seen to be of poor quality. There are a number
of sources for the poor quality data: incorrect subtraction of bright stars, scattered sunlight, and
corrupt data rolls (Toller 1981). The poor quality data were removed from consideration using
four criteria. First, data contaminated with scattered sunlight (data taken within 70 ffi of the sun
for Pioneer 10 and within 45 ffi for Pioneer 11) were removed. Second, all data with negative values
were removed as these were the result of interuptions in the datastream of between the spacecraft
and ground station. Third, the data were divided into 5 ffi \Theta 5 ffi boxes and data inside each box
deviating over 3 standard deviations from the average in either their blue measurements, red
measurements, or red/blue ratio were removed. Fourth, a small number of points were removed
by visual inspection. Approximately 25% of the IPP measurements were of poor quality.
The resulting good data were used as the input for the algorithm described above to produce
the final high­resolution maps. The algorithm was iterated 10 times, until little change was
detected. The best iteration map to use depended on the region being investigated. For low

-- 8 --
Table 1. IPP Usable Days
Spacecraft year day R a
[years] [days] [AU]
Pioneer 10 1972 354 3.26
Pioneer 10 1973 149 4.22
Pioneer 10 1973 237 4.64
Pioneer 10 1973 279 4.81
Pioneer 11 1974 57 3.50
Pioneer 10 1974 68 5.15
Pioneer 11 1974 106 3.81
Pioneer 11 1974 148 4.06
Pioneer 11 1974 178 4.22
Pioneer 11 1974 236 4.51
Pioneer 11 1974 267 4.66
a Sun­spacecraft distance, R, taken
from NSSDC WWW pages.

-- 9 --
Fig. 3.--- The Aitoff projection (galactic longitude of zero in the center) of the Pioneer blue image of
the sky is displayed. The resolution of the displayed map is 0: ffi 5 by 0: ffi 5. The large hole corresponds
to the Sun's location as seen from Pioneer 10/11. The intensity units are S 10 (V) G2V .

-- 10 --
galactic latitude regions where the amplitude of real structure in the maps is much larger than the
noise amplitude, the 10th iteration gave the best map. For high galactic latitude regions where the
real structure amplitude is smaller than the noise amplitude, the 1st iteration gave the best map.
The 10th iteration map for the blue is displayed in Figure 3. The 10th iteration red map is
similar to the blue map. Typical uncertainties were 2% for the blue and 3% for the red. Figure 3
can be compared directly with the blue background as seen by the Hipparcos star mapper. A map
of this background is presented in Figure 6 of Wicenec & van Leeuwen (1995). The comparison
is good both in overall strength and morphology. From this comparison, the uniqueness of the
Pioneer maps was quite apparent as the Pioneer blue map lacks the substantial zodiacal light seen
in the Hipparcos star mapper blue map.
As we were only concerned with high latitude regions, we will use the 1st iteration map for
the rest of this paper and save the higher iterations for later work. The 1st iteration blue and red
maps are just smoother versions of the 10th iteration maps.
3. Photometric Star and Galaxy Counts
In order for this investigation to succeed, the contribution to the Pioneer blue and red
measurements from stars and galaxies fainter than m V = 6:5 was needed. We have tackled this
problem by constructing a Master Catalog from three separate catalogs, each complete in a subset
of the range between 6.5 and ¸20th magnitude. Ironically, the stars and galaxies with magnitudes
between 12 and ¸20 (Palomar O & E) have the best photometric data available due to the
existence of the Automated Plate Scanner Catalog of the Palomar Sky Survey I (APS Catalog,
Pennington et al. 1993). In the magnitude range between 9 and 15 (¸V band in the north and
¸B in the south), the Guide Star Catalog (GSC, Lasker et al. 1990; Russell et al. 1990; Jenkner
et al. 1990) provides data in only one band. For the magnitude range between 6.5 and 9.5, there
exists no good complete photometric catalog. We have used a combination of catalogs (see x3.2)
to construct a Not So Bright Star Catalog (NSBS Catalog) to give the best currently available
positions and magnitudes for stars with magnitudes between 6.5 and 9.5.
3.1. Transformations Between Photometric Systems
Underlying the construction of the Master Catalog was the transformation between the
Palomar blue & red (O & E) magnitudes and Johnson B & R magnitudes to Pioneer blue &
red (PB & PR) magnitudes. The normalized response curves, R(–), for the blue and red bands
of all three photometric systems are shown in Figure 4 (Minkowsi & Abell 1963; Lamla 1982;
Toller 1981). The Palomar blue (O) response function was computed for an airmass of 1.5 (Hayes
& Latham 1975) in order to reproduce the transformation between the Palomar and Johnson
systems used in calibrating the APS Catalog (Humphreys et al. 1991). For all 6 above bands as

-- 11 --
Fig. 4.--- The response curves are plotted for the Pioneer blue and red channels (PB & PR, Toller
1981), the Johnson B and R (Lamla 1982) and the Palomar blue and red (O & E, Minkowsi &
Abell 1963).

-- 12 --
well as the Johnson V band, the band's equivalent wavelength (– eq ), equivalent bandpass (\Delta– eq ),
zero magnitude flux (F – ), and the intensity corresponding to S 10 (V) G2V and S 10 (V) A0V units were
computed and are tabulated in Table 2. The values of – eq and \Delta– eq are obtained by
– eq =
R –R(–) d–
R
R(–) d–
(4)
and
\Delta– eq =
R R(–) d–
R max
; (5)
respectively. The flux corresponding to a magnitude of zero, F – , in each band was calculated by
summing the product of the band's response curve and a calibrated spectrum of ff Lyrae (T¨ug,
White, & Lockwood 1977). The calibrated spectrum of ff Lyrae was multiplied by 1.028 before
use to account for the fact that ff Lyrae's V magnitude is 0.03 (Hoffleit & Warren 1991). The
intensity corresponding to one S 10 (V) G2V unit and one S 10 (V) A0V unit was computed by summing
the product of each band's response curve with the spectrum, set to 10th magnitude in the V
band, of the sun (Lockwood, T¨ug, & White 1992) and ff Lyrae (T¨ug, White, & Lockwood 1977),
respectively. One S 10 (V) X unit is defined as intensity equivalent to one 10th V magnitude star
of spectral type X per square degree where X is either G2V or A0V. The intensity in mag/ut 00
corresponding to S 10 (V) G2V units in the B bands, V band, and R bands is 28.5, 27.8, and 27.5
mag/ut 00 , respectively.
The transformations from the Palomar and Johnson systems to the Pioneer system were
accomplished by means of the above band response functions (Figure 4) and zero magnitude fluxes
(Table 2) along with an observational grid of stellar spectra spanning the Hertzsprung­Russell
diagram (Silva & Cornell 1992). This grid consists of spectra covering 3510­8930 š A with a
resolution of 11 š A and includes 72 spectral types spanning spectral classes O--M and luminosity
classes I--V. Most of the spectra are for solar metallicity stars, but some are for metal­rich and
metal­poor stars. The spectra were dereddened and stars of similar spectral types were averaged
to produce the final 72 spectral type spectra (Silva & Cornell 1992).
In order to check the accuracy of our transformations, the transformation from the Johnson
system to the Palomar system was computed and compared to the same transformation as
determined by Humphreys et al. (1991). Figure 5a displays the (O \Gamma B) correction as a function of
(B \Gamma V ) which transforms the Johnson B magnitude to the corresponding Palomar O magnitude.
Figure 5b displays the (E \Gamma R) correction as a function of (V \Gamma R) which transforms the Johnson
R magnitude to the corresponding Palomar E magnitude. The agreement between the (O \Gamma B)
and (E \Gamma R) corrections derived in this paper and those of Humphreys et al. (1991), validates this
method for deriving transformations between photometric systems.
The transformation from the Palomar to the Pioneer system is displayed in Figure 6 and
the transformation from the Johnson to the Pioneer system is shown in Figure 7. We have
fitted the resulting curves with polynomial functions in order to have an analytic form for the
transformations. The number of terms in the fitted polynomial was determined by adding terms

-- 13 --
Table 2. Photometric Band Details
System Band – eq \Delta– eq F –
a S 10 (V) G2V S 10 (V) A0V
[ š A] [ š A] [ergs cm \Gamma2 s \Gamma1 š A \Gamma1 ] [ergs cm \Gamma2 s \Gamma1 š A \Gamma1 sr \Gamma1 ]
Johnson B 4467 1014 6.632\Theta10 \Gamma9 1.198\Theta10 \Gamma9 2.174\Theta10 \Gamma9
Palomar O 4249 1168 6.343\Theta10 \Gamma9 1.087\Theta10 \Gamma9 2.080\Theta10 \Gamma9
Pioneer PB 4370 826 6.997\Theta10 \Gamma9 1.192\Theta10 \Gamma9 2.294\Theta10 \Gamma9
Johnson V 5553 881 3.639\Theta10 \Gamma9 1.193\Theta10 \Gamma9 1.193\Theta10 \Gamma9
Johnson R 6926 2057 1.950\Theta10 \Gamma9 8.813\Theta10 \Gamma10 6.394\Theta10 \Gamma10
Palomar E 6412 386 2.289\Theta10 \Gamma9 9.828\Theta10 \Gamma10 7.505\Theta10 \Gamma10
Pioneer PR 6441 968 2.305\Theta10 \Gamma9 9.919\Theta10 \Gamma10 7.558\Theta10 \Gamma10
a F – is the flux corresponding to a magnitude of zero. See text for details.
Fig. 5.--- The transformation from the Johnson system to the Palomar system is plotted. The
(O \Gamma B) correction is displayed in (a) and the (E \Gamma R) correction is displayed in (b). Note that the
(O \Gamma B) and (E \Gamma R) corrections derived in this paper agree quite well with those from Humphreys
et al. (1991).

-- 14 --
Fig. 6.--- The transformation from the Palomar system to the Pioneer system is plotted. The
(PB \Gamma O) correction is displayed in (a) and the (PR \Gamma E) correction is displayed in (b). The
maximum corrections for both (PB \Gamma O) and (PR \Gamma E) are small.
Fig. 7.--- The transformation from the Johnson system to the Pioneer system is plotted. The
(PB \Gamma B) correction is displayed in (a) and the (PR \Gamma R) correction is displayed in (b). While the
maximum correction for (PB \Gamma B) is small, the maximum correction for (PR \Gamma R) is large due to
the significantly different values of – eq for the PR and R response curves.

-- 15 --
until the resulting fitted polynomial followed the general trend of the points. The maximum
corrections for (PB \Gamma O), (PR \Gamma E), (PB \Gamma B), and (PR \Gamma R) are ¸0.25, ¸0.10, ¸0.20, and ¸1.0,
respectively. The large maximum correction for (PR \Gamma R) is due to the large difference in the – eq
value between the Pioneer (6441 š A) and Johnson (6926 š A) systems.
3.2. Master Catalog Construction
Three star and galaxy cat