�������� ���� �� ���� ��������� ������. ����� ������������� ��������� : http://www.mso.anu.edu.au/~bessell/Mags/Mags.ps ���� ���������: Thu Jan 22 02:20:30 2004 ���� ��������������: Tue Oct 2 00:48:03 2012 ���������: ��������� �����: � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �

Magnitude Scales and Pho�
tometric Systems
More than any other aspects of astronomy, the
subjects of magnitude scales and photometric
systems are encumbered by history. The in�
tensity of light from stars and other cosmic
objects is usually expressed in magnitudes, an
inverse logarithmic scale that confuses physi�
cists who work in SI units, but that is practi�
cal for astronomers.The apparent magnitude of
an object is a measure of the intensity of ra�
diation within a particular wavelength interval
received from that object at the Earth. The
absolute magnitude is the magnitude that the
object would have were it situated at a distance
of 10 parsecs [(pc), about 32.6 light�years (ly)]
from the Sun. The relation between apparent
magnitude m and absolute magnitude M is m
� M = 5 1og d � 5, where the distance d is in
parsecs.
The total energy, integrated over all wave�
lengths, received at the Earth from an object is
also expressed as a magnitude, the bolometric
magnitude. The difference between the bolo�
metric magnitude M bol , and the magnitude mA
in bandpass A is called the bolometric correc�
tion BCA . BC without a qualifier normally
refers to the correction to the visual magnitude;
BC, or BC V . The zeropoint of the bolomet�
ric magnitude scale is usually set by adopting
M bol (sun) = 4.75, which infers a BC sun =�0.07.
Early astronomers compared star with star,
a procedure that still retains great benefits.
The surface temperatures of common stars
range from 30,000 K down to 3000 K, and
their apparent brightnesses cover a range of
almost a factor of 10 10 , from the sky back�
ground upwards (this range does not include
the Sun). The majority of stars are constant
in total light output and in temperature and
as no laboratory lamps have energy distribu�
tions very similar to those observed in stars it
is natural that astronomers seek to use the stan�
dard candles in the sky. Photometric systems
represent attempts to define standard band�
passes and sets of standard sources, measured
with these bandpasses, that are well�distributed
about the whole sky. Different photometric sys�
tems measure different wavelength bands. All
photometric systems enable the measurement
of relative fluxes, from which can be inferred
particular properties (such as temperature) of
the emitting object, but different systems claim
to do it more precisely or more efficiently than
other systems. Some of the systems were de�
veloped and modified by different astronomers
over many years and the literature contains
confusing versions and calibrations. Some peo�
ple have despaired that it is too confusing and
have suggested that we should start again with
a well�defined ultimate system, but recent anal�
ysis has shown that modern versions of the ex�
isting photometric systems can be placed on a
firm quantitative basis and that more care with
passband matching will ensure that precise and
astrophysically valid data can be derived from
existing, though imperfect, systems.
Several recent large scale astronomical
projects are providing significant new magni�
tudes for large numbers of a wide range of astro�
nomical objects. Firstly, the gravitational lens�
ing projects, such as MACHO, EROS, OGLE
and AGAPE are identifying a range of variable
stars and measuring light curves in the Galac�
tic Bulge and the Magellanic Clouds; secondly,
wide field survey projects, such as the Sloan
Digital Sky Survey are measuring the mag�
nitudes of all objects above a certain bright�
ness in the northern sky. Finally, the remark�
able astrometric satellite HIPPARCOS mea�
sured extremely precise magnitudes for more
than 118000 stars, mostly brighter than 9th
magnitude, over the whole sky and from outside
the atmosphere. These projects are indicative
of the quality and quantity of data that is be�
coming available and that will be a great chal�
lenge for standard magnitude calibrations.
1

The Magnitude Scale of Hipparchus; the
Intensity Scale of Pogson
In the earliest recorded star catalog, Hip�
parchus (second century BC) divided the stars
in the sky into six groups. Twenty of the
brightest stars that could be seen were called
first�magnitude stars and those at the limit
of visibility were called sixth�magnitude stars.
Intermediate�brightness stars were put in inter�
mediate magnitude classes. In the eighteenth
century, astronomers were using telescopes and
had begun to measure the light intensities of
stars by closing down the telescope aperture
until the image of the star under study just
disappeared (the disappearance aperture). By
taking the ratio of the squares of the disappear�
ance apertures of two different stars, the rela�
tive intensity of the stars' light could be cal�
culated. This was the beginning of astronomi�
cal visual photometry. Norman R. Pogson (in
1856) at the Radcliffe Observatory compared
his measurements of stellar brightness with stel�
lar magnitudes given in contemporary star cat�
alogs (such as those of Stephen Groombridge
and the zone observations of Friedrich Arge�
lander and Friedrich Wilhelm Bessel) and sug�
gested the simple relationship m = 5 log a +
9.2 to relate the magnitude m of a star and the
disappearance aperture a (in inches). This rela�
tion implies a coefficient of 2.5 for the relation
between magnitude and the logarithm of the
intensity as I / a 2 . Around that time, Gustav
Theodor Fechner and Wilhelm Edward Weber
(1859) were investigating the response of the
eye to light and proposed the following psy�
chophysical law: m � m 0
= s log I/I 0
, where m
is a perceived brightness and the constant s de�
fines the scale. Pogson's work implied a scale of
� 2.50 for astronomical visual (eye) photometry,
and we thereby have the basis for the inverse
logarithmic scale. There was continuing dis�
agreement concerning the adoption of this exact
scale of 2.5 and it was not until almost 30 years
later, after the Harvard photometry was pub�
lished in 1884, that adoption was assured.The
constant m 0
, which defines the zero point, has
undergone much refinement since Pogson's es�
timate and was officially set by the specified
visual magnitudes of stars in the ''north po�
lar sequence.'' The early photometry catalogs
are based on this sequence but the magnitude
scale today is established by the contemporary
whole�sky photometric standard star catalogs.
New Magnitudes from New
Detectors
Technological advances over the last 100 years
have provided a series of light detectors to sup�
plement the eye. These detectors, in general,
respond differently to light of different wave�
lengths than does the eye; that is, they are
more sensitive to blue light or to red light than
is the eye. The advent of photography in the
late nineteenth century revolutionized astron�
omy, as did the introduction of photomultiplier
tubes with their light�sensitive photocathodes
in the mid�twentieth century and sensors such
as silicon charge�coupled devices (CCDs) and
infrared detectors over the last 20 years. Light
intensities, or magnitudes, measured with these
new detectors naturally differ from the visual
magnitudes and depend on the color of the
star. Initially, there was only the difference
between visual magnitudes and ''blue'' photo�
graphic magnitudes to be considered, but sev�
eral factors resulted in a proliferation of differ�
ent passbands and photometric systems: the
extension of photographic and photocathode
sensitivities to a wider wavelength range; the
use of colored glass filters and interference fil�
ters to sample the starlight in narrower bands
within the total wavelength sensitivity range of
the detectors and more recently, the require�
ments of survey instruments to provide maxi�
mum sensitivity for faint object detection.
2

Rationale for Multicolor Photom�
etry
Much photometry of astronomical objects is
carried out in order to measure the apparent
total brightness of objects and their relative
brightnesses at different wavelengths, that is,
their energy distributions. It is possible to
characterize the temperatures of most objects
from the overall shapes of their energy distribu�
tions. It is also possible to infer the metal con�
tent of stars from depressions in their energy
distributions at particular wavelengths. These
depressions (absorption lines) are due to the
absorption of flux, principally by Fe and Ti
(which have very rich line spectra), Ca, Mg,
and the molecules CN and CH (which have
very strong lines in the blue� violet region of
cool stars). There are many other molecular
absorption bands (such as TiO, CO, and H 2 0)
that depress the continuum in very cool stars;
such molecular features are also used to pro�
vide information on the temperature, chemical
composition, and luminosity. The energy dis�
tributions of galaxies and star clusters can be
analyzed to extract the relative numbers of dif�
ferent kinds of stars making up these composite
objects. Redshifts of very distant galaxies and
quasi�stellar objects can also be measured from
the positions of depressions or peaks in their
energy distributions. These are called photo�
metric redshifts and have been used very suc�
cessfully with data from the Hubble Space Tele�
scope and the Sloan Digital Sky Survey.
Multicolor photometry is best thought of
as very�low�dispersion spectroscopy. The en�
tire high�resolution spectrum of a star or other
cosmic object contains a large amount of infor�
mation, but when dealing with extremely faint
objects or with large numbers of objects, it is
a great advantage to measure a small number
of wavelength bands in as short a time as pos�
sible. Such a minimal technique is invaluable if
it enables the derivation of many of the same
parameters obtainable from a complete (and
very redundant) description of the spectrum. A
great deal of effort therefore has gone into accu�
rately measuring and calibrating colors and de�
pressions in terms of temperatures, metal abun�
dances, and other parameters, and investigat�
ing which of competing minimal descriptions of
a star's spectrum is the most accurate or most
practical.
Finally, gravitational lensing results in
changes in the brightness of the object
independently of the color measured. Most in�
trinsically variable stars however, have differ�
ent amplitudes in red and blue light. Conse�
quently, gravitational microlensing surveys are
efficiently carried by splitting the light between
a blue and a red channel for simultaneous di�
rect comparison. Equal amplitude variations in
blue and red channels imply a lensing event not
a variable star.
Photometric Systems: Nat�
ural and Standard
A light detector, a telescope, a set of filters,
and a method of correcting for atmospheric ex�
tinction make up a natural photometric system.
Each observer therefore has their own natural
system. The standard system is indirectly de�
fined by a list of standard magnitudes and col�
ors that have been measured for a set of typical
stars, using the natural system of the origina�
tor. These are often called the primary stan�
dards. Later lists comprising more stars and
fainter stars but based on the primary stan�
dards are called secondary standards. However,
in the case of all photometric systems, recently
published secondary standards effectively rede�
fine the standard system because they tend to
be more accurately measured than the primary
lists and to represent contemporary detectors,
filters, and practice.
The term ''color'' is an abbreviation for
color index, which is the difference between the
apparent magnitudes in two different spectral
regions. Photometry has generally been pub�
lished as a series of colors and a single magni�
3

tude. The zero points of many color systems
are set so that ff Lyrae (Vega) has zero colors.
In the southern hemisphere (where Vega is in�
accessible) and often also in the north, the zero
point is set by requiring that an ensemble of
unreddened A0 stars have colors of zero mag�
nitude.
The Original Standard Sys�
tems
The most influential of the early works of
photoelectric photometry were the broadband
Johnson UBVRI and Kron RI systems, which
covered the wavelength region between 310 and
900 nm (3100 and 9000 � A). The natural sys�
tems of Harold L. Johnson and of Gerald E.
Kron and co�workers served as ''standard'' sys�
tems for many other users who attempted with
varying success (due to differences in detectors,
filters, telescopes, and techniques) to duplicate
the originators' natural systems. That is, us�
ing their own detectors and filters, astronomers
measured stars from the Johnson and Kron lists
and linearly transformed their natural magni�
tudes and colors to be the same as the Johnson
and Kron colors and magnitudes. They then
applied those same linear coefficients to trans�
form the colors and magnitudes of unknown
stars onto the Johnson or Kron system. The
original blue and yellow filters were chosen by
Johnson from readily available glasses so that
when used with the 1P21 photo�multiplier tube
they approximated the ordinary blue (B) pho�
tographic response (� 436 nm) and the visual
(V) response (� 545 nm). A more violet magni�
tude U (� 367 nm), which is useful for very hot
stars, was obtained by using a common violet
glass. In retrospect, these choices should have
been based more on astrophysics and less on
glass availability, but so much work has been
done in this UBV system that the weight of
history assured its continuation. Intercompar�
ison of much of the published broadband pho�
tometry (in particular, photometry taken more
than 15 years ago) often shows scatter of more
than 0.03 magnitudes, but more recent photom�
etry obtained using better equipment, better
matched natural systems, and better secondary
standard stars agrees to better than 0.01 mag�
nitude, or 1%.
The 1P21 phototube was a remarkable in�
vention and its high blue sensitivity dominated
the development of photometric systems for
over 30 years.There were red�sensitive devices
available but observations were made only for
bright stars because for many years these de�
vices were much less sensitive, noisier and less
reliable than the 1P21. In the mid�1970s new
detector materials became available; in partic�
ular, the gallium�arsenide and multialkali pho�
totubes, which provided high (AE 15% quan�
tum efficiency) sensitivity between 300 and 860
nm, and the infrared�sensitive InSb (indium an�
timonide) photodiodes together with low�noise
preamplifiers, which revolutionized photometry
between 1000 and 4000 nm. Both developments
enabled photometry to be done on faint stars
that had hitherto been the sole province of the
blue�sensitive detectors. Photometry done with
the new red�sensitive tubes was placed on ei�
ther the Kron or the Johnson standard system,
again with mixed success, and it has only been
in the last 10 years that A. W. J. Cousins'
RI ''near�natural'' standard system (based on
the Kron system) has gained widespread ac�
ceptance. It has also been very useful that the
Cousins system's R (� 638 nm) and I (� 797
nm) bands are similar to the contemporary
photographic R and I bands.
Johnson also introduced the infrared alpha�
betic JKLMN (approximately 1.22, 2.19, 3.45,
4.75, and 10.4 �m) system in the mid�1960s, us�
ing PbS (lead sulfide) detectors and bolometers.
The water vapor in the Earth's atmosphere de�
fines a series of wavelength bands (windows)
through which observations from the ground
can be made, Johnson used interference filters
(and, unfortunately, the atmospheric H 2 0 ab�
sorption bands) to define what he called the J,
K, L, M, and N bands. Ian S. Glass, in his
4

early observations with an InSb detector, used
the additional band H (� 1:63�m) between I
and K, and in his choice of filters attempted to
match the other Johnson bands. All infrared
observers have proceeded in a similar fashion
and have concentrated mainly on copying the
Johnson K magnitude scale. Identical detectors
have been used but a range of slightly differ�
ent filters and observatory altitudes have pro�
duced subtly different systems. The publica�
tion of sufficient numbers of stars in common
from the different natural systems has helped
delineate the differences, and transformations
between the systems are now quite reliable.
Passbands or Response Functions
The most important specifications of a photo�
metric system are the passbands or response
functions of its magnitudes. For a variety of
reasons, technical and historical, the passbands
of the original broadband photometric systems
have not been known with certainty and this
has inhibited close matching of natural sys�
tems and has prevented computation of accu�
rate synthetic colors from theoretical spectra.
The recent availability of spectrophotometry
for many stars combined with the increased pre�
cision of second�generation photometric cata�
logs has, however, enabled the passbands to be
derived indirectly by computing synthetic col�
ors from spectrophotometry of stars with well�
defined standard colors and adjusting the pass�
bands until the computed and standard cata�
log colors agree. This technique has enabled
the passbands of the major systems to be well�
defined, which in turn has permitted filters to
be designed that still will result in good pass�
band matches with a variety of detectors. In
addition, when it is not possible to match pass�
bands exactly with some detectors, such as pho�
tographic plates, it is possible to predict ac�
curately the differences between photographic
and photoelectric magnitudes by computing the
synthetic magnitudes using the different pass�
bands.
In Fig.I the normalized passbands of the
Johnson�Cousins UBVRI system are shown
plus an added Z band for CCD observations.
The F � (flux per unit frequency interval) spec�
trum of an A0 star is shown for orientation.
Table 1 lists the effective wavelengths � eff , the
approximate bandwidth \Delta�, which is the full
width at half maximum of the passband, and
the absolute calibration of the UBVRIJHKL
system, based on the flux of Vega, for a zero�
magnitude A0 star. Note that the effective
wavelengths of the broad bands change with the
color of the objects. The effective wavelengths
listed are for an A0 star.
Other Photometric Systems
Real or perceived drawbacks in existing photo�
metric systems (the UBV system in particular)
stimulated the design of other photometric sys�
tems better suited for measuring temperatures,
metal�line blanketing, effective gravity, and in�
terstellar reddening. Some of these systems
used broad bands comparable to the UBVRI
system, while others used narrower bands de�
fined by different mixes of glass filters or inter�
ference filters. Effective wavelengths and other
specifications of some of the better�known sys�
tems are given in Table 2 and are discussed be�
low.
Geneva and Walraven Systems
Difficulties with matching natural systems have
been eliminated by the strategy employed by
proponents of the Geneva (UBB 1
B 2
VV 1
G) and
Walraven (VBLUW) systems. The latter takes
its name from Th. and J. H. Walraven. These
multiband photometric systems are supervised
by small groups who control the instrumenta�
tion and supervise the data reduction and cali�
bration. The colors have been well calibrated in
terms of gravity, temperature, and abundance.
Such closed systems have excellent precision,
but not necessarily greater than that possible
5

from the open Cousins UBVRI system with
careful bandpass matching.
Washington System
This CMT 1
T 2
system was devised to use the
wideband sensitivity of the extended�red de�
tectors, to improve the sensitivity of blue� vio�
let colors to metallicity and gather more violet
light in cool stars and to try to separate the
effects of CN from other metal lines. We have
found that the violet C band is a very useful
metallicity indicator for faint K giants but that
the M band contains little more information
than does V; T 1
and T 2
, have no advantages
over R and I. The minimal CVI system is very
useful for metal�weak K stars.
Str�omgren Four�Color System
The uvby system was devised by Bengt
Str�omgren to measure better the Balmer dis�
continuity, the metallicity, and the temperature
of A, B, and F stars. The bands are essentially
separate unlike the UBV bands, which overlap.
The u band is completely below the Balmer
jump; v measures the flux near 400 nm, a region
with much absorption due to metal hnes; b is
centered near 460 nm and is affected much less
than B by metal�line blanketing; y is essentially
a narrower V band. The u filter is colored glass,
the others are interference filters. Two special
indices are derived: m 1
= (v \Gamma b) \Gamma (b \Gamma y), which
measures metallicity, and c 1
= (u \Gamma v) \Gamma (v \Gamma y),
which measures the Balmer discontinuity. The
index (b � y), like B � V, is used primarily as a
temperature indicator. The system is capable
of very high precision but, unfortunately, er�
rors in the width of v filters manufactured some
years ago resulted in nonstandard filters being
supplied to many users. Since then, published
photometry has exhibited some systematic dif�
ferences in c 1
, and m 1
, and there are difficulties
in computing these colors from theoretical spec�
tra, particularly for cool stars. Recent standard
catalogs of new and more homogeneous obser�
vations are of high precision and internal con�
sistency and it should now be possible to define
better the v band. Two additional interference
filters (15 and 3 nm wide) centered on the Hfi
line are often used together with the four col�
ors. The Hfi index is used to derive Iuminosi�
ties in B stars and reddening in F and G stars.
The Str�omgren system was the first photomet�
ric system devised to measure specific stellar
features. Because of the short wavelength base�
line of its four color filters, 1% photometry at
least is required to utilize the system's advan�
tages over the UBVRI system.
DDO (35,38,41,42,45,48) System
This system (also built around the sensitivity
of the 1P21 photomultiplier) was designed for
the analysis of G and K dwarfs and giants. The
35 filter is the u filter of the four�color system;
the 38 filter is also a glass filter and better mea�
sures metal blanketing than the v filter, being
further to the violet and wider; 41 measures
the CN band; 42, 45, and 48 are continuum fil�
ters.The color 35�38 (the 3538 index) measures
the Balmer jump, 3842 measures the metallic�
ity, and 4245 and 4548 are used for gravity and
temperature measurements. By restricting the
measurements to the blue spectral region, com�
plicated corrections for spectral line blanketing
are necessary to derive temperatures and gravi�
ties. Good results, especially for faint K dwarfs,
can be obtained by using V � I or R � I as
the temperature indicator. Because of the nar�
row bandwidth of some of the filters, the DDO
(David Dunlap Observatory) system has been
mainly restricted to relatively bright stars.
Thuan�Gunn System
The uvgr system of Trinh Xuan Thuan and
James E. Gunn was devised in the mid�1970s
from the UBVR system for use with an S20
photocathode detector and in order to avoid the
strong mercury emission lines from city lights
and [OI] lines in the night sky. The g and r
bands are of similar width to the V and R bands
whereas the u and v bands are about half the
6

width of the U and B bands. The g � r color
has a longer baseline than V � R but transforms
well.
Photographic Systems
Originally photographic emulsions were only
sensitive to light blueward of 490 nm. These
were the O emulsions. Different chemical sen�
sitizing shifted the red sensitivity cutoff to
longer wavelengths: G 580 nm; D 650 nm; F
700 nm; N 880 nm, approximately. By using
blue�cutoff glass filters and the red cutoff of
the emulsions, various photographic passbands
were made. Photographic U used a violet fil�
ter for both blue and red cutoffs. Attempts
were made to convert the photographic colors
onto the photoelectric UBVR system but these
were not often very accurate because of lim�
itations in iris photometry and poor matches
of the bandpasses. In recent years, astronom�
ical photography has undergone a renaissance
caused, first, by the development of new fine�
grain emulsions (Kodak IIIaJ, IIIaF and more
recently TechPan) and the utilization of meth�
ods of greatly increasing the sensitivities of the
J, F and TechPan emulsions (using hydrogen
gas) and of the N emulsions (using silver ni�
trate solution) and, second, by the use of new
scanning microdensitometers and better meth�
ods of intensity calibration. Averages of sev�
eral wide�field Schmidt camera plates or higher
scale prime�focus plates can now produce pho�
tometry to a few percent to very faint limits.
Theoretical investigation of bandpasses enables
better filter design for bandpass matching or
predicts the relevant transformations and sys�
tematic differences between photoelectric and
photographic photometry. Photographic pho�
tometry these days is usually restricted to at�
tempted matches to the Johnson U and B or the
Thuan�Gunn g systems using IIIaJ plates, to
Cousins R or Thuan�Gunn r using IIIaF plates,
and to Cousins I using IVN plates. Direct pho�
tographic calibration from step�wedges is usu�
ally supplemented by direct magnitude mea�
surements of stars in each field using a CCD
array.
CCD Photometric Systems
The high quantum efficiency of CCDs and their
inherent linearity have made them the detec�
tors of choice in recent years for most areas of
photometry. Unfortunately, the advantages of
the CCDs were initially not fully attained be�
cause some users paid insufficient care to define
their passbands and to standardizing their pho�
tometry. This resulted in internally precise re�
sults but an inability to relate these results with
much confidence to the standard system data
or to theoretically derived magnitudes and col�
ors. Astronomers now realize the importance
of matching their CCD passbands to standard
passbands or deriving accurate passbands for
their natural systems to enable them to be cal�
ibrated using synthetic photometry.
The standard BVRI system can easily be
realised with thinned CCDs and colored glass
filters but the U system is more problematical
due to the lower UV response of many CCDs.
The Z band, between I and J is also now often
added to CCD UBVRI based systems.
The HST WFPC2 Photometric
System
The Wide Field Planetary Camera (WFPC2)
on the Hubble Space Telescope has a suite of in�
terference filters that cover both the space�UV
and the optical spectrum. Although not identi�
cal to the well established UBVRI and uvby
systems, there are passbands that are quite
similar. A lot of attention has been given to
calibrating the WFPC2 system both from ac�
tual ground based observations where possible
and from synthetic photometry so that excel�
lent transformations are possible to the older
standard systems and reliable temperature cal�
ibrations can be made from model atmosphere
fluxes. Fig 2 shows some of the WFPC2 band�
7

passes. Table 3 lists effective wavelengths (for
an A0 and M0 star) and FWHM of some of the
HST passbands and those of the other CCD
systems discussed below.
The HIPPARCOS and TYCHO
Photometric Systems
The ESA Astrometric Satellite, Hipparcos,
used three independent photometric detectors.
The main Hipparcos passband (H p ) corre�
sponded primarily to the spectral response of
the S20 photocathode of the image dissector
scanner combined with the transmission of the
optics. The Tycho photometric data were de�
rived from the star trackers and measured mag�
nitudes in B T and V T , with passbands some�
what similar to standard B and V. Fig 3 shows
the Hipparcos and Tycho passbands in relation
to the standard BVR passbands. The large
width of the H p passband results in significant
systematic differences between the H p magni�
tudes and standard V magnitudes, depending
on reddening, metallicity and luminosity. Nev�
ertheless, the extremely high precision H p mag�
nitudes (� 0:0015) and the lower but still good
precision (� 0:012) for the Tycho V T magni�
tudes combined with the whole�sky coverage
makes these catalogs an invaluable resource,
not only for measurements of individual stars,
but for enabling intercomparisons to be made
between and within ground based photometric
systems.
The Sloan Digital Sky Survey
Photometric System
This photometric system comprises five color
bands (u 0 , g 0 , r 0 , i 0 and z 0 ) that divide the entire
CCD sensitivity range between the atmospheric
UV cutoff near 300 nm and the CCD cutoff near
1100 nm. The passbands, related to those of the
Thuan�Gunn system, and shown in Fig 4, are
essentially nonoverlapping and most are wider
than those of the UBVRI system ensuring high
efficiency for faint object detection. For ease of
transformation into other systems or duplica�
tion of the system by others, it would have been
better were the bands to have overlapped more
with a less rectangular profile but the system
itself will be very well defined by observations
made with a duplicate detector and filters on
a separate telescope. Unlike most other photo�
metric systems, the zeropoints of the SDSS sys�
tem have been placed on the spectrophotomet�
ric AB magnitude system defined by the abso�
lute fluxes of four F subdwarfs. The passbands
are essentially filter defined and in general have
blue edges defined by a colored glass and red
edges by a short pass interference coating (see
article on filters).
Gravitational Lensing Projects
The MACHO Photometric System
The MACHO project of monitoring for gravi�
tational microlensing events utilizes simultane�
ous CCD imaging in two passbands by sharing
the light between two cameras using a dichroic
beam splitter. The blue and red bands are fur�
ther limited using an interference filter on the
red side and the sensitivity cutoff of the thick
CCDs on the blue side. The blue band ap�
proximate a broad blue shifted V band while
the red band approximates the R band. Good
transformations are possible to the VR system
and reliable calibrations are possible using syn�
thetic photometry. As well as detecting many
microlensing events, the Macho project has pro�
vided unique and invaluable data on variable
stars in the Magellanic Clouds and the Galac�
tic Bulge.
The EROS Photometric System
The EROS1 observations were taken consecu�
tively through two broad band filters BE and
RE that produced respectively, passbands mid�
way between B and V and R and I. Two differ�
ent sets of filters were used during the course of
the observations and the BE \Gamma RE colors were
transformed into V�I. EROS2, like MACHO,
8

Table 1. Johnson�Cousins�Glass UBVRIJHKLM System
U B V R I J H K L M
� eff (nm) 367 436 545 638 797 1220 1630 2190 3450 4750
\Delta� (nm) 66 94 85 160 149 213 307 39 472 460
F � (V = 0)
(10 \Gamma30 Wcm \Gamma2 Hz \Gamma1 ) 1780 4000 3600 3060 2420 1570 1020 636 281 154
Table 2. Effective Wavelengths (nm) and FWHM Bandpasses (nm) for Selected Photoelectric
Systems
� eff \Delta� � eff \Delta� � eff \Delta�
Geneva U 350 47 Walraven W 323.3 15.4 Washington C 391 110
B 424 76 U 361.6 22.8 M 509 105
B 1 402 38 L 383.5 21.9 T 1 633 80
B 2
448 41 B 427.7 49.0 T 2
805 150
V 551 67 V 540.6 70.3
V 1
541 44
G 578 47
Str�omgren u 349 30 DDO 35 349.0 38.3 Thuan�Gunn u 353 40
v 411 19 38 381.5 33.0 v 398 40
b 467 18 41 416.6 8.3 g 493 70
y 547 23 42 425.7 7.3 r 655 90
fi w 489 15 45 451.7 7.6
fi n 486 3 48 488.6 18.6
has two cameras and the light is divided using
a dichroic beamsplitter. The division is made
at a redder wavelength (� 650 nm) than for the
EROS1 system and the EROS2 blue band more
resembles the Hipparcos passband.
9

Table 3. Effective Wavelengths (nm) and FWHM Bandpasses (nm) for CCD Based Systems
� eff � eff \Delta� � eff � eff \Delta� � eff � eff \Delta�
HST 336 334 339 47 HIP B T 421 439 70 SDSS u 0 356 360 64
439 430 443 71 V T 526 542 100 g 0 475 500 135
450 451 477 107 H P 517 595 230 r 0 620 632 137
555 532 559 147 i 0 761 772 154
675 667 675 127 UBV B 436 464 94 z 0 907 907 147
814 788 805 147 V 545 558 85 EROS BE1 485 506 109
R 641 666 160 RE1 657 679 191
MACHO B 519 543 144 I 791 799 143 BE2 539 575 190
R 682 700 178 Z 909 906 96 RE2 767 796 260
10

Figure captions
Fig 1. The passbands of the standard UBVRI
system.
Fig 2. Some of the passbands of the HST
photometric system.
Fig 3. The Hipparcos passbands in com�
parison with the standard BVR passbands. H P
is shown by the thickest line; B T and V T the
medium thick lines and BVR the thin lines.
Fig 4. The Sloan Digital Sky Survey pass�
bands.
MICHAEL BESSELL
11