Author's personal copy
Journal of Cultural Heritage 12 (2011) 54–64
Original article
An evaluation of daylight distribution as an initial preventive conservation
measure at two Smithsonian Institution Museums, Washington DC, USA
Julio M. del Hoyo-Meléndeza,∗,b, Marion F. Mecklenburga, María Teresa Doménech-Carbób
a Museum Conservation Institute, Smithsonian Institution, 4210 Silver Hill Road, Suitland, MD 20746-2863, USA
b Instituto Universitario de Restauración del Patrimonio, Universidad Politécnica de Valencia, Camino de Vera 14, s/n (edificio 9B), 46022 Valencia Espan˜a, Spain
a r t i c l e i n f o
Article history:
Received 13 May 2009
Accepted 25 May 2010
Available online 25 June 2010
Keywords:
Illuminance measurements
Natural lighting
Data logging
Museum environment
Smithsonian American Art Museum
National Portrait Gallery
a b s t r a c t
This paper presents the results of a light levels survey conducted at the Donald W. Reynolds Center for
American Art and Portraiture in Washington DC. The museum space is shared by the National Portrait
Gallery and the Smithsonian American Art Museum. After six years of extensive renovations, the building
reopened to the public in July 1, 2006. The structure was not originally designed to house a museum
collection since it contains numerous openings such as windows, doors and skylights, which provide a
path for natural radiation to enter the building and come in contact with the artworks. From a preventive
conservation standpoint, this is an important problem since sensitive works of art in the collection may
be subjected to damage caused by light exposure. Environmental data loggers installed throughout the
museum were programmed to take successive measurements every 10min for 24h a day, 7days a week
and 52weeks a year. This light levels assessment started in November 1, 2007 and finished in October
31, 2008. This study presents a new method for determining natural radiation exposures registered in
exhibition spaces that rely on both electric lighting and natural lighting, considering the growing trend
of using daylight illumination in museums.
© 2010 Elsevier Masson SAS. All rights reserved.
1. Research aims
The main objective of the study was to evaluate the relative
distribution of light throughout different exhibition areas of the
Donald W.Reynolds Center for American Art and Portraiture, dur-
ing a light levels survey of one year. The initial goal of the research
was to establish and develop a measurement methodology that
helped to assess natural and artificial lighting contributions in
spaces where both types of radiation are employed. The inves-
tigations also had the intention of verifying the effectiveness of
current preventive conservation measures used to minimize the
solar illumination reaching the interior spaces of the building.
This study forms part of a comprehensive preventive conser-
vation program, which includes various lines of research. First,
light-fastness surveys of museum materials using micro-fading
spectrometry have been conducted jointly with the aim of investi-
gating the exposures capable of inducing noticeable color changes.
Second, an investigation of the reciprocity principle of light expo-
sures has also been performed with the objective of identifying
deviations, which are dependent on light-stability of the mate-
∗ Corresponding author. Tel.: +1 301 238 1283; fax: +1 301 238 3709.
E-mail addresses: judeho1@doctor.upv.es (J.M. del Hoyo-Meléndez),
mecklenburgm@si.edu (M.F. Mecklenburg), tdomenec@crbc.upv.es
(M.T. Doménech-Carbó).
rial. Third, different illumination ratios employed throughout the
museum have allowed devising a method for estimating the pho-
tometric exposures received by objects on display. This procedure
takes into account the higher illuminance levels experienced by the
object due to the use of direct light sources in addition to indirect
sources. The final goal is the establishment of lighting exhibition
guidelines at the two museums based on results gathered during
the previously described research campaigns.
2. Introduction
Museum personnel often face the challenge of lighting and pre-
serving objects on display since there should be a balance between
pleasing viewers with adequate light levels that permit correct
appreciation of artifacts, while addressing the possible amount of
energy striking the object’s surface. Natural lighting is often pre-
ferredoverelectric lightingwhen illuminatingworksof art, because
the first one provides a more pleasing rendering effect due to its
spectral power distribution. However, natural lighting has a large
component of UV radiation, which is known for its detrimental
effects to the artwork. Therefore, this type of radiation needs to
be minimized. Museums worldwide have traditionally used the
maximumrecommended levels of illuminationproposedby Thom-
son [1]. The difficulty relies on applying these guidelines without
conducting a systematic light monitoring program, which helps to
1296-2074/$ – see front matter © 2010 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.culher.2010.05.003
Author's personal copy
J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64 55
Table 1
Summary of various recommended annual exposure limits for susceptible museum artifacts.
Recommending organization Low sensitivity (klxh/y) Moderate sensitivity (klxh/y) High sensitivity (klxh/y)
Illuminating Engineering Society of North America Variable 480 50
Heritage Collections Council, Australia – 507–650 127–200
Canadian Conservation Institute – 1000 (ISO 4)a 100 (ISO 2)
Commission Internationale de l’Éclairage 600 150 15
a These values are based on light-fastness ratings established by the International Standard Organization (ISO).
qualitatively and quantitatively evaluate the visible light and UV
radiation present in museum exhibition spaces. The spectral distri-
butionof the light source is an important aspect of this kindof study
since similar illuminance levels may contain different amounts of
UV radiation depending on the source. The source with a higher
UV component, usually natural lighting, is more damaging than its
equivalent electric source. Current accepted levels of UV radiation
are 35 and 75microwatts per lumen (Wlm−1) for highly sensitive
and moderately sensitive objects, respectively [2]. Every museum
preventive conservation policy must take into account the sensi-
tivity to light of artists’ materials found in their collections before
establishing adequate light levels for their objects.
The importance of measuring the light levels in museum build-
ings due to potential detrimental effects to sensitive works of
art on display has been indicated by several authors [3–5]. Some
researchers [6,7] have incorporated annual photometric exposure
limits, for works of art on paper, which are based on the maxi-
mum recommended levels published earlier by Thomson. Various
authorities in the museum lighting research field have published a
series of guidelines which include annual exposure limits for sus-
ceptible artifacts [8–11]. A summary of these recommendations is
presented in Table 1. These maximum recommended values are
based on International Standard Organization (ISO) light-fastness
classifications. These ratings are then converted to exposures based
on available data on the fading of typical museum materials.
Moreover, annual exposures can be extrapolated to determine the
approximate number of years that a material would require to
experience a perceptible color change. The differences in maxi-
mum allowable exposures observed in Table 1 are mainly due to
the number of hours per day and days per year in which these val-
ues are based. For example, amaximumannual exposure of 50klxh
for high sensitive materials recommended by the Illuminating and
Engineering Society of North America (IESNA) is based on illumi-
nation at 50 lx, for 8h in a day over a period of 125days in a year. In
contrast, the Canadian Conservation Institute (CCI) recommends a
limit of 100klxhy−1 for a fugitive material that belongs to the ISO
2 class. This value is based on 8h per day and 250days per year,
at the same illumination level. In the CCI scale, an ISO 2 material
displayed for 250days per year is expected to experience a per-
ceptible color change after 10years (1.0Mlxh exposure). The same
type of material exhibited for 125days per year, as specified in the
IESNA guideline, would take twice as long to reach the same degree
of damage. A guideline proposed by the Commission Internationale
de l’Eclairage (CIE) in2004divides thematerials into four categories
based on their responsiveness to visible light. Although initially the
CIE standard may seem too restrictive, the scales become equiva-
lent after carefully examining the types of materials belonging to
each category. For example, materials such as oil paintings classi-
fied as moderately sensitive in the three other scales belong to the
low responsivity group in the CIE system. Therefore, a comparison
of the CIE and IESNA guidelines reveals that the maximum rec-
ommended exposures for oil paintings are 600 and 480klxhy−1,
respectively.
Numerous investigations on the interaction of light with
museum materials [12–14] and the use of dosimeters to estimate
photometric exposures in museums [15,16] have been conducted.
Bacci et al. have developed a new kind of light dosimeter, which
consistedof amixtureof photosensitivedyes andapolymer applied
on apaper substrate [16]. In this paper, the authors accuratelymen-
tion some of the problems related to measuring light levels in the
museum, which can be divided into three principal areas. First,
the difficulty of measuring radiation coming from mixed sources
(natural and artificial) and subjected to seasonal changes. Sec-
ond, extrapolations of pointmeasurements oftenprovideuncertain
cumulative photometric exposures. Third, placement of an illumi-
nance meter next to each object is unmanageable and results in
elevated costs for the institution. Despite all of these aspects, a
few studies concentrate on proposing a systematic methodology
for evaluating natural and artificial radiation levels and presenting
results derived from field measurements [17,18].
For all the above-mentioned reasons, this work presents a
new promising methodology for cultural institutions interested in
quantifying the distribution of natural radiation in their exhibi-
tion spaces. This study includes three major innovative aspects.
First, photometric and UV radiation readings were recorded con-
tinuously every 10min for an entire year. This has provided the
opportunity of examining separately the data from any desired
interval of a day, month or year. Second, the illuminance data
presented parallel with the cumulative exposure curves offers an
original way of evaluating the results, since daylight behaviors can
be easily inferred from the combination charts. Third, this new
method offers the possibility of separating the radiation into its
natural and artificial components. This separation is based on the
assumption of a constant artificial lighting baseline in illuminance
and a discernable natural radiation signal at specific times of the
day. Continuous measurements are important since they permit
to evaluate changes throughout the day and those taking place
at different times of the year. Comparable exhibition areas can
be evaluated by determining the natural lighting distribution of
one area, which then allows extrapolating the results to similar
locations within the building. Finally, data loggers provide a more
accurate method for calculating total photometric exposures when
compared to estimations obtained from light dosimeters.
2.1. The Old Patent Office Building
The Old Patent Office Building is the third public structure con-
structed in early Washington DC after the White House and the
United States Capitol. It was designed by American architect Robert
Mills in the Greek Revival style characteristic of the 19th century
in the United States [19]. The construction of the building begun
in 1836 and it was completed in 1868. The Old Patent Office Build-
ing has witnessed many important historical events throughout
its history and it was employed by several government agencies
before it was converted into a national museum. The building is a
National Historic Landmark and was included in the National Reg-
ister of Historic Places of the United States of America in 1966. Two
major renovations took place during the 20th century, resulting in
a transformation of the space from office spaces into exhibition
galleries. After the most recent six-year renovation, the building
reopened to the public in July 1, 2006, andwas renamedTheDonald
W.ReynoldsCenter forAmericanArt andPortraiture. This historical
Author's personal copy
56 J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64
Fig. 1. Schematic description of the proposed method.
structure covers an entire city block demarcated by F and GStreets
and 7th and 9th Streets NW in Chinatown, Washington DC. The
building consists of four wings, which surround a central court-
yard that has an undulating glass roof. Every wing has an external
portico and a central corridor, with galleries distributed along both
sides. The galleries are either adjacent to the street or next to the
central courtyard and have been defined as external and internal,
respectively.
3. Methodology
An overview of the methodology employed in this study is
shown in Fig. 1. Potential users should follow this outlined pro-
cedure if interested in implementing the methodology in other
museums or cultural institutions around the world. Initially, it is
necessary to conduct an evaluation of the building and its exhi-
bition spaces. The building site in terms of its geographic location,
climatic zone andweather patterns needs to be investigated. A rep-
resentative number of areas must be selected taking into account
the different display areas within the building. Some of the aspects
that need to be considered regarding themonitoring zones include:
floor area, fenestration, ceiling height, construction materials, out-
door obstructions and type of works exhibited in the space. The
different zones illuminated with natural daylight within a selected
space must be defined next. For example, various orientations such
as south-facing versus east-facing walls or horizontal in opposition
to vertical surfaces are specified at this stage. Afterward, the test
points are chosen for each defined zone. In the present study, the
selection of light monitor locations took into consideration several
architectural aspects such as four building wings, three different
floors, internal and external galleries, horizontal and vertical sur-
faces and various window treatments, among others. The objective
was to assure that these measuring points would be representative
of several lighting situations encountered throughout the build-
ing. At this point, various aspects such as measuring devices, data
acquisition system, monitoring schedule and frequency of mea-
surements are selected. For example, the sampling interval selected
for this study was one reading every 10minutes; measurements
were then logged automatically and saved into a single file. A total
of 24monitors were installed with various positions towards the
sun. The distribution included 19monitors mounted to walls next
to the artwork. Three meters were horizontally positioned near
three-dimensional works of art located under skylights or near
windows. Two meters were placed inside display cases that were
directly below skylights on the west wing. A building plan that
includes the location and orientation of 12meters installed on the
second floor is shown in Fig. 2. Data loggers were named using
the following format: building wing initial, floor number and a
two-digit gallery number. In special cases where two meters were
employed in the same gallery, initials such as EFW, WFW, SFW or H
were added to designate an east-facing wall, a west-facing wall, a
south-facingwall or a horizontal surface, respectively. For example,
meter E112 EFW was located in the east wing, first floor, gallery 12
and was mounted to an east-facing wall.
After all these aspects were considered, the measurement cam-
paignwas implemented. The datawas then downloaded, evaluated
and presented to staff from both museums on a monthly basis in
order to identify problematic areas and to provide an opportunity
for making adjustments during the course of the study. Validation
of the method is performed using a computer simulation program,
whichhas alreadybeen validated. This allows comparing simulated
and measured illuminance data from winter and summer months
acquired for various museum locations. This computer simulation
program allows estimating the amount of daylight reaching an
exhibition space taking into account the geographic location of
the building, the various seasons of the year and different sky
illuminance models. If there is good agreement between calculated
and measured data, the method is validated and further treatment
of the illuminance data is possible. Numerical integration is
employed for obtaining cumulative photometric exposure curves
Author's personal copy
J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64 57
Fig. 2. Second floor plan of the museum showing approximate locations and orientations of 12 light monitors (image courtesy of Smithsonian Institution).
and consequently calculating the total exposure. Even though
contributions from both natural and artificial radiation during
most of the day are evident, the average artificial lighting baseline
of each gallery provides a way for estimating the artificial lighting
exposure, since this type of radiation is easier to be controlled
than its natural counterpart and its intensity is roughly constant.
Hence, the artificial radiation exposure is obtained by multiplying
the number of electric lighting hours times the artificial lighting
baseline in illuminance. The assessment of illuminance data
permits to accurately establish the lighting schedule and the
average baseline produced by electric lights. In the present study,
an additive effect of natural radiation to the artificial lighting com-
ponent was noticed after inspecting the illuminance data. These
high illuminance values of natural lighting resulted in daylight
signals, which exceeded the artificial lighting baseline during a
specific time interval. Therefore, the natural lighting exposure was
obtained by calculating the area under the daylight curve, after
subtracting the artificial lighting baseline from the data.
After concluding the measurement campaign, the annual natu-
ral daylight performance of the spaces is evaluated. The different
annual trends in daylight distribution are identified at this stage
and also the areas of high risk of damage for the objects. The results
are then contrasted with maximum recommended museum expo-
sures taking into consideration the light-sensitivity of thedisplayed
works, the museum opening hours and the artificial lighting com-
ponent registered in each space. The data can be eventually used
towards the establishment of museum lighting and exhibition pro-
tocols. The results will have an influence on decisions such as the
use of a larger or lower natural lighting component in a space and
the frequency of rotation of exhibitions.
Thismethod has repercussion onmuseumenergy savings appli-
cations since it permits to identify exhibition areas where a lower
artificial lighting baseline can be employed while introducing a
higher natural lighting component; especially in spaces where
objects with moderate or low light-sensitivity are displayed. The
method also finds application in the museum lighting design field.
The production of visually pleasing exhibits often requires illumi-
nation using a large natural lighting component, which permits
better color rendering of the artworks. This methodology would
allow assessing individual lighting scenarios in order to determine
if an increment in natural lighting component is feasible. Never-
theless, the integration of the method in preventive conservation
programs in museums is probably the most important application.
Knowing the natural daylight distribution in a room throughout
the year permits better control through the introduction of win-
dowtreatmentsat specific locationsduringaparticular seasonorby
reducingunnecessary layers during themonthswhennatural light-
ing poses a lower risk. This work also lays the foundation for future
studies in other institutions located in different climatic zoneswith
different solar radiation amounts and distribution.
3.1. The geographical location of the museum and its relationship
to sun paths
As previously stated, it is important to get acquainted with
the geographical coordinates of the museum site relative to
the International Meridian. Geographic coordinates for the Don-
ald W.Reynolds Center are latitude 38◦53′52.5′′ N and longitude
77◦01′21.90′′ W and were obtained using EZ-locate online geocod-
ing software [20]. This information is essential when searching
Author's personal copy
58 J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64
Fig. 3. Astronomy and geography related concepts: (a) the various sun’s paths across the celestial sphere as experienced by an arbitrary building. The dashed lines represent
the equinoxes and solstices (adapted from E.C. Boes [21]); (b) change in total sunshine hours with latitudes, the position of the museum site is indicated by a discontinuous
line (adapted from G.N. Tiwari [22]).
for solar and meteorological information available at various web-
based databases. The earth’s inclination of 23.5◦ with respect to
its rotational axis and its rotation around the sun give rise to the
seasons, due to higher or lower proximity of the hemispheres to
the sun. The pre-Copernican model of the sun rotating around the
earth is typically used to simplify the discussion about sun paths
[21]. An arbitrary building located in the Northern Hemisphere and
covered by the celestial sphere has been used to describe sun paths
at various times of the year. This model can be used to track inci-
dent sunrays on the sphere that can ultimately make their way
into the building. As viewed from the earth, the sun is higher in
the sky during the summer and it reaches its lowest altitude dur-
ing the winter. In Fig. 3a are shown the resulting sun trajectories
that give rise to four solar phases namely vernal equinox, sum-
mer solstice, autumnal equinox and winter solstice. The variation
of total number of sunshine hourswith latitudes is shown in Fig. 3b.
The data shows that there are 12h of natural daylight in March
and September since declinations are approximately 0◦ at all lati-
tudes [22]. It can also be observed that sunlight hours will increase
or decrease at higher latitudes, depending on which solstice takes
place. The 38◦53′52.5′′ north latitude of the Donald W.Reynolds
Center is indicatedbyadiscontinuous line that intersects the curves
at approximately 9.5h and 15h of sunlight for December and June
solstices, respectively. These sunshine hours have a clear effect
on the monthly natural radiation exposures, confirming that this
type of evaluation is essential prior to starting any monitoring
program.
Natural daylight enters the museum through side windows,
glass doors and skylights. Fig. 4a shows the museum’s Great Hall
viewed from the third floor showing skylights on the south wing.
Twoexamples ofmeter set-ups arepresented: thefirst case showsa
monitor thatwas vertically installed (Fig. 4b) next to anoil painting.
The second example shows a meter that was horizontally mounted
(Fig. 4c) on a sculpture’s pedestal. The method of reducing UV
radiation and visible light levels is by using 3MTM ScotchtintTM
NV-15 Night Vision window film, as well as either Thermoveil®
shade cloth or Venetian blind window treatments employed by the
Smithsonian American Art Museum and National Portrait Gallery,
respectively.
3.2. Instrumentation
The meters used in this survey were ELSEC 764C environmen-
tal monitors manufactured by Littlemore Scientific Engineering
(Dorset, United Kingdom). These monitors allowed to take periodic
readings of the amount of visible light in lux (lmm−2) and the UV
component inWlm−1. Themonitor is cosine corrected for angular
response and the radiation is detected by a pair of silicon photodi-
odes connected to a single chip microprocessor. UV radiation and
visible light wavelength ranges are 300–400nm and 400–700nm,
respectively. The accuracy of the ELSEC 764C instrument for visi-
ble light and UV radiation readings is 5% and 15%, correspondingly.
Monitorswere installedonall threefloors anddistributed in the fol-
lowing manner: four on the first floor, 12 on the second floor and
eight on the third floor. The meters were programmed to take mea-
surements every 10min, for 24h a day, 7days a week and 52weeks
a year. A USB2IR external wireless mini adapter from StarTech.com
was employed for instantaneous transfer of files from the monitors
to a portable computer. The data was downloaded using RView 3.8
for Loggers from Littlemore Scientific Engineering, then converted
into a spreadsheet for data reduction and preparation of plots.
3.3. Calculation of total, artificial and natural photometric
exposures
Light data was organized into a series of plots of change in illu-
minance with time. The total exposure, HT, in lux-hours (lx h) was
obtained by calculating the area under the illuminance curve using
numerical integration. The integral was approximated by adding
the areas of all rectangles obtained after dividing the total region
into 10-minute intervals. The integral can be expressed by Eq. (1),
where Ev is the illuminance integrated over the exposure time, t.
HT =
∫ t
0
Evdt = 12
n−1∑
i=1
(Evi + Evi+1 )(ti+1 − ti) (1)
Each monitor recorded a set of n points: (t1, Ev1), (t2, Ev2),. . .,(tn,
Evn), every 10min. These points define n−1 rectangular intervals,
where the width of the i-th interval is given by the difference
between two successive measurements, as shown in Eq. (2).
ti = ti+1 − ti (2)
The height of the rectangle associated with this i-th interval is
the average illuminance at that specific interval shown in Eq. (3).
Evi =
Evi + Evi+1
2
(3)
Assuming a constant artificial lighting baseline in illuminance
and absence of natural radiation, the artificial lighting exposure,
HA, was obtainedbymultiplying the average artificial lighting value
times the duration of exposure, as shown in Eq. (4).
HA = Ev × t (4)
Author's personal copy
J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64 59
Afterward, the natural lighting exposure, HN, was obtained by
subtracting the artificial lighting exposure from the total exposure
as shown in Eq. (5).
HN = HT − HA (5)
Monthly natural lighting exposures and cumulative exposure
curves were plotted on a single graph using a primary and a sec-
ondaryy-axis, respectively. Furthermore, thegeneratedcumulative
exposure curves served to evaluate the resulting daylight illumina-
tion trends.
3.4. Method validation
The method has been validated using Superlite 2.0 [23]. This
software allows simulating the natural daylight distribution in a
room by calculating hourly values of illuminance on any given
plane that describes the space. The calculation technique used is
based on the radiation flux exchange between surfaces by divid-
ing the examined space into a mesh of small elements. Interior
natural daylight levels for any sun and sky condition can be mod-
eled for spaces having windows or skylights. The program takes
into account the variation of the fenestration transmittance with
the angle of incidence and also the effect of the direct solar
radiation and the reflected light component. Some of the param-
eters considered include the dimensions of the space, the overall
fenestration area, the reflectance and the transmittance of the
construction materials, and the orientation of display surfaces.
Hourly indoor illuminance levels were simulated from 9:00h to
16:00h in various display areas of the museum during winter and
summer days with the aim of comparing expected and measured
values.
Measured and Superlite-simulated indoor illuminance in a
south-facing test-point of gallery S222 are shown in Fig. 5a. This
room has a floor area of 64m2 and it is side-lighted by two win-
dows located on the south wall, each one with an area of 2.4m2.
In December 20, under clear sky conditions, the illuminance reg-
istered by meter S222 SFW varied between 2 and 47 lx while the
simulated illuminance for an equivalent surface ranged from 15 to
44 lx. The same surface exhibited subtle changes throughout the
day under overcast sky conditions observed in December 13, with
measured and simulated values that ranged from 2 to 5 lx and 1
to 4 lx, respectively. Simulated and measured illuminance trends
obtained for two summer days were fairly similar and remained
within the 3 to 12 lx range. A reasonably low discrepancy was
observed in June 13, which consisted of a slower increase in illu-
minance recorded between 10:00h and 12:00h relative to the
predicted augment. Equivalent south-facing display areas located
in the first, second and third floor also showed good agreement
between simulated and measured data. Measured and simulated
indoor illuminance values for a horizontal surface of gallery S321
are shown in Fig. 5b. This space has a floor area of 494m2 and
it is mainly illuminated by a large skylight that has an area of
187m2. Simulated and measured illuminance data for Decem-
ber 20 (clear sky) showed similar natural daylight patterns which
fluctuated from 49 to 183 lx and 39 to 243 lx, respectively. In con-
trast, overcast sky conditions observed in December 13 resulted
in less variation in illuminance throughout the day, which ranged
from 6 to 23 lx and 13 to 47 lx for simulated and measured data,
respectively. For all examined cases, the differences in absolute
illuminance values ranged from 5% to 25%. The data shows a fairly
good agreement in both range and distribution pattern consider-
ing the difficulty of accurately simulating the daylight availability
in thebuilding. Simulated andmeasured illuminancedata is consis-
tent even under various sky conditions and at different times of the
year.
4. Experimental data and results
An example showing visible light values recorded by one of
the monitors on a single day is included to illustrate how the
Fig. 4. (a) the Great Hall viewed from the third floor mezzanine showing skylights on the south wing of the building; (b) light monitor E112 EFW vertically installed near an
oil painting; (c) light monitor S321 H horizontally installed on a sculpture pedestal. (Photos: Julio M. del Hoyo-Meléndez).
Author's personal copy
60 J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64
Fig. 5. Measured versus Superlite-simulated indoor illuminance in twoof the exhibition spaces examined: (a) south-facing test-point of gallery S222; (b) horizontal test-point
of gallery S321.
data was treated. An intense signal produced by natural lighting
and a constant artificial lighting baseline can be deduced from the
data. After acquiring the data, the calculations described in Section
3.3 were performed. The same procedure was used for deter-
mining monthly and yearly natural lighting exposures for every
monitor.
4.1. Determination of the natural lighting exposure using the
proposed model
Illuminance data recorded by monitor E252 SFW in September
7, 2008, is shown in Fig. 6. This meter was located in a southeast
Fig. 6. Illuminancedata (primaryaxis) recordedbymonitor E252 SFWinSeptember
7, 2008, and the cumulative exposure (secondary axis) obtained from numerical
integration. A gray horizontal line is used to indicate the artificial lighting baseline
in illuminance.
gallery on the second floor and was mounted to a south-facing
wall. Sunrise and sunset times on that day were 5:43h and 18:28h,
respectively. The data show a gradual increase in illuminance from
about 6:30h until 8:40h, when gallery lights were switched on.
This is indicated by an abrupt increase in illuminance pointed out
by the arrow in Fig. 6. An illuminance maximum of 185 lx, primar-
ily due to the effect of natural lighting, was observed at 9:10h.
This was followed by a decrease in illuminance observed from
9:20h to 13:00h. Artificial lighting clearly had a dominating effect
from 13:00h to 19:30h, indicated by an average artificial light-
ing baseline of 9.3 lx. This was deduced from the data recorded
between 18:40h and 19:40h, time when only artificial lighting
was detected. The lights were then switched off from 19:50h to
00:00h, hence the gallery had an artificial lighting baseline of 9.3 lx
for an 11-hour period. Assuming that the illuminance produced by
the electric light sources was constant, an artificial lighting expo-
sure of 0.1 klxh is obtained, bymultiplying the average illuminance
times the number of hours of electric lighting. The total exposure
obtained for 11h was 0.3 klxh, which gave an estimated natural
lighting exposure of 0.2 klxh. This simple example helps to illus-
trate how natural and artificial lighting contributions to the total
exposure can be quantified as 67% and 33%, respectively. Even
though the two components of light were present from 8:40h
to 19:50h, the data shows that natural lighting had a significant
additive effect from 8:40h until 18:30h. This is characterized by
daylight data with illuminance values higher than the 9.3 lx arti-
ficial lighting baseline. The effect of natural lighting was observed
to a lower extent from 13:00h until 18:30h. This was indicated by
a slightly higher average illuminance of 10.7 lx produced by sub-
sequent positions of solar rays with increasing distance from the
meter location.
Author's personal copy
J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64 61
Fig. 7. (a) overall monthly natural lighting exposures obtained for 23monitors; (b) detail of the 0–25klxh range showing monitors divided into four groups according to
their calculated exposures.
4.2. Monthly natural lighting exposures and cumulative exposure
curves
Fig. 7a shows the monthly natural lighting exposures recorded
by each monitor over the entire year. The same graph expanded
along the 0–20klxh exposure range is presented in Fig. 7b. It
can be observed that 17 vertically installed monitors located
in side-lighted spaces registered lower monthly natural light-
ing exposures, which were typically below 5klxh. In this group,
third floor monitors experienced natural lighting exposures that
were approximately two times higher than the ones recorded
in the first and second floor. Two horizontally installed meters
located in the second floor registered natural lighting exposures,
which were approximately three times higher than the ones
recorded by their vertical counterparts. The highest monthly natu-
ral lighting exposures were registered by four monitors installed
below skylights. The monitors were divided into four groups
based on calculated annual natural lighting exposures. Group I
consists of vertically installed monitors located in the first and
second floor in side-lighted spaces. The natural lighting expo-
sures registered by these 13monitors ranged from5 to 33klxhy−1.
Four vertically installed monitors also located in side-lighted
galleries in the third floor constitute the second group. These
meters recorded natural lighting exposures that ranged from 38
to 52klxhy−1. Two horizontally installed monitors located in
the second floor form the third group since they recorded nat-
ural lighting exposures of 104 and 116klxhy−1. Finally, four
monitors installed directly under skylights located in the third
floor comprise the fourth group. This later group can be fur-
ther divided in two classes: fully exposed meters and monitors
inside display cases. For instance, fully exposed monitors located
in gallery S321 registered 398 and 1451klxhy−1 and the dif-
ferences were due to their placement in vertical and horizontal
surfaces, respectively. In contrast, monitors placed inside dis-
play cases experienced similar annual exposures of 173 and
175klxh.
Themonitors belonging to groups I and IImeasured illuminance
levels on vertical surfaces in which objects with moderate to high
light-sensitivity are exhibited. The artifacts displayed in these gal-
leries range frompaper documents to oil paintings and their annual
natural lighting exposures averaged 15 and 42klxhy−1 for group I
and II, respectively. Although these yearly exposures remain under
themaximumrecommendedvalues, it is important to also take into
account the electric lighting component detected in these galleries
which was usually in the 90–150klxhy−1 range. Thus, materials
with moderate sensitivity would generally receive total annual
exposures that are under the maximum recommended levels in
all four scales described in Table 1. However, the exposure limits
for highly sensitive materials proposed by the CCI and the Her-
itage Collections Council (HCC) of Australia seem more adequate
for comparison with the present study since the two examined
museums are open 10h a day for 364days per year. For instance,
any object exhibiting high light-sensitivity located in an exhibition
area that belongs to either group I or II could have experienced an
exposure in the 105–165 or 130–170klxhy−1 range, respectively.
Group III monitors measured the amount of radiation striking two
different horizontal surfaces adjacent to various wooden furniture
pieces identified as moderately light-sensitive. The calculated total
exposures for these twometerswere in the200–250klxhy−1 range
and therefore remained under the maximum recommended limits.
According to the CCI scale, these two objects could experience a
perceptible color differencewithin 40–50years of continuous exhi-
bition after reaching a total exposure of 10Mlxh. Finally, objects
exhibited in group IV zones have moderate to low light-sensitivity.
Artifacts exhibited in gallery S321 and inside display cases located
in the west wing consist of oil and acrylic paintings as well as
inorganic-based objects such as stone and metal sculptures. The
annual artificial lighting components registered in gallery S321
were approximately 22 and 225klxh for a horizontal and a ver-
tical surface, respectively. Therefore, the horizontal surfaces of
the more permanent artifacts experienced total annual exposures,
whichwere in the1500klxhvicinity. Although this valuemay seem
relatively high, it is not of great concern since these objects are
composed of permanent materials. In contrast, vertical surfaces in
gallery S321 registered total exposures, which were approximately
600klxhy−1. This value remains within the annual exposure lim-
its for oil paintings except for the IESNA guideline for moderately
susceptible materials, which recommends 20% less. Finally, meters
placed inside display cases located in the third floor recorded total
annual exposures, which were in the 300–400klxh range. Thus,
materials exhibited in these areas remained within the maximum
exposure limits for moderately sensitive materials according to the
four-museum lighting guidelines used for comparison.
The procedure described in the previous section was applied
to every meter at the end of each month. An example of monthly
natural lighting exposure and cumulative exposure curve of one
year obtained for meter E252 SFW is shown in Fig. 8. This meter
registered an increase in natural lighting exposure from Decem-
ber 2007 until March 2008, when a maximum exposure of 3.8 klx
was observed. A decrease in exposure was then observed from
March 2008 until May 2008, followed by a new rapid increase
from July 2008 until September 2008, when a new maximum
Author's personal copy
62 J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64
Fig. 8. Monthly natural lighting exposures and cumulative exposure curve obtained
formeter E252 SFWshowingexposuremaxima inMarch2008andSeptember2008;
therefore its cumulative exposure exhibits Type A behavior.
exposure of 3.9 klx was registered. The natural radiation expo-
sure for November 2007was obtained using numerical integration.
Afterward, the December 2007 exposure was determined simi-
larly and added to the exposure obtained for the previous month
to calculate the cumulative exposure for these two months. The
same procedure was successively applied to obtain the cumu-
lative exposure for the following months. A line connecting
all cumulative exposure points permits to evaluate the annual
trends in daylight illumination. Location and orientation of the
meter gave rise to two exposure maxima observed during the
months of vernal and autumnal equinoxes when solar declina-
tions were −0.02◦and 0.17◦, respectively. All meters vertically
installed in south- and west-wing galleries and two horizontally
installed meters located on side-lighted spaces in the second
floor exhibited a natural lighting behavior similar to the one
described in Fig. 8, identified as type A for the purpose of this
paper.
A different trend in cumulative exposure was observed for
meters located in side-lighted vertical zones in the north and east
wings aswell as for the fourmeters locatedbeneath skylights on the
third floor. This behaviorwas identified as TypeB just for differenti-
ation purposes. For example, cumulative exposure curves obtained
for meters E111 WFW and E112 EFW show gradual increases from
December 2007 until June 2008, when maximum natural lighting
exposures of 2.6 klxh and 1.9 klxh were registered (Fig. 9). This
was followed by decreases in natural lighting amount, recorded
from June 2008 until October 2008. For one year, the total natural
lighting exposures recorded by meters E111 WFW and E112 EFW
were 19klxh and 10klxh, respectively. This 9 klxh difference is
explained by meter locations on external and internal east-wing
galleries. The meter on an external gallery was mounted to an east-
facingwall and therefore received a higher natural lighting amount
than its counterpart, which was installed on a west-facing wall at
an adjacent internal gallery.
Fig. 9. Monthly natural lighting exposures and cumulative exposure curves
obtained for meters E111 WFW and E112 EFW showing exposure maxima in June
2008; therefore their cumulative exposures exhibit Type B behavior.
These two natural lighting distributions can also be distin-
guished in Fig. 7b inwhich overallmonthly exposures for allmeters
are shown. The two behaviors observed are consistent with data
calculated for the city of Dublin by McVeigh [24]. Although Dublin
is located at higher north latitude than Washington DC, it was also
possible to identify two natural lighting behaviors, which were
dependent on orientation and ratio of vertical to horizontal radia-
tion. In the cited study, solar radiant exposure measurements were
given in MJm−2 but they can be somehow correlated to photomet-
ric measurements, since the two quantities are proportional. Solar
radiant exposure is a radiometric quantity used to describe and
quantify the irradiation level on a surface considering UV radiation,
visible light and infrared radiation. In contrast, photometric expo-
sure based on illuminance measurements exclusively indicates the
amount of radiation spectrally weighted by the standard photo-
metric visibility curve, which has a maximum at the wavelength of
555nm for the human eye.
A summary of the results is presented in Table 2. Some of the
general trends observed included higher exposures registered in
external galleries relative to the internal ones and also higher expo-
sures recorded by horizontal meters relative to the ones vertically
installed. The data also indicated that natural lighting exposures
increase as one moves up from the first floor to the third floor.
East-wing galleries typically received higher exposures than those
located on the south wing; this effect was noticeable after eval-
uating the data recorded by monitors located in the first floor
and third floor. As discussed earlier, larger natural radiation values
were recorded by meters located under skylights on the third floor.
Therefore, objectswithmoderate tohigh light-stability exhibited in
this area were recommended to be periodically rotated, taking into
account that the highest natural lighting exposures are expected in
June, around the summer solstice.
Table 2
Summary of natural lighting distributions observed for various museum locations.
Museum floor Type A behavior (groupa) Type B behavior (group)
First Side-lighted vertical surfaces in the south wing (I) Side-lighted vertical surfaces in the east wing (I)
Second Side-lighted vertical surfaces in the west and south wings (I) Side-lighted vertical surfaces in the north and east wings (I)
Side-lighted horizontal surfaces in the north and south wings (III)
Third Side-lighted vertical surfaces in the south wing (II) Side-lighted vertical surfaces in the east wing (II)
Top-lighted vertical and horizontal surfaces below skylights (IV)
a These numbers are used to classify the exhibition areas, which were divided into four groups based on annual natural lighting exposures obtained.
Author's personal copy
J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64 63
Table 3
Comparison of UV radiation data and total photometric exposures recorded by five
monitors in December 2007.
Monitor UV radiation (Wlm−1) Total natural light exposure (klxh)
S234 SFW 338 18
E252 H 30 7.8
S321 SFW 28 9.7
S321 H 0 28
S222 SFW 0 3.4
4.3. UV radiation data
Small amounts of UV radiation, derived from solar illumina-
tion, were detected by four meters namely S234 SFW, E252 H,
S321 SFW and S321 H. The meter located in gallery S234 was an
exceptional case since it was used to measure unfiltered sunlight
striking a south-facing wall of a south-wing gallery in December
2007. Architectural glass elements in this gallery had no treat-
ment against natural radiation and this served to estimate how
much protection the current window treatments provide. These
results are presented in Table 3. Meters S234 SFW and S222 SFW
registered UV radiation averages of 338Wlm−1 and 0Wlm−1,
respectively. These twometersweremounted to south-facingwalls
on south-wing galleries and recorded total natural radiation expo-
sures of 18 and 3.4 klxh. This indicates that window treatments
employed in gallery S222 were accountable for an 81% reduc-
tion in natural lighting exposure relative to the one registered in
gallery S234. A 100% reduction in UV radiation was noticed after
comparing the data from these two galleries. The two horizontal
meters, E252 H and S321 H, registered a natural radiation expo-
sure variation of approximately 20klxh. This was most likely due
to the location of the latter one in a space illuminated using a sky-
light. However, no UV radiation was registered by meter S321 H in
December 2007 probably due to the low solar altitude observed
near the winter solstice. Meters E252 H, S321 SFW and S321 H
registered UV radiation yearly averages of 24, 23 and 7Wlm−1,
respectively. These averages remain under the maximum recom-
mended levels of 30Wlm−1 and 75Wlm−1, for very sensitive
and sensitive objects [2].
4.4. Action of natural and artificial lighting on the artworks
Illuminancemeasurements taken near the surface of the objects
have revealed that artifacts may receive higher exposures due to
the use of direct sources, which are sometimes employed as accent
lights. In absence of natural radiation, it is possible to determine the
direct to indirect artificial lighting ratio. These spot measurements
can be used in conjunction with continuous readings recorded by
a fixed monitor adjacent to the artwork, with the aim of evaluating
the individual effects of natural and artificial lighting through-
out the year. Preliminary results indicate that in cases where a
direct artificial lighting source has a dominating effect over natural
radiation, the annual exposure receivedbyanobjectmaybeextrap-
olated from a spot measurement recorded near its surface, since
the contribution from natural radiation is almost negligible. This
effect has been observed in galleries in which the natural radiation
component is significantly reduced, either by the use of win-
dow treatments or due to their location adjacent to tall evergreen
trees planted on the exterior of the building. In contrast, increases
in diffuse daylight signal observed at various times of the year
may result in a significant contribution of natural radiation to the
total exposure. As a consequence, a higher exposure than the one
extrapolated from a single photometric measurement is generally
obtained. This phenomenonhas been identified in variousmuseum
exhibition spaces containing skylights. The results have allowed
making more informed decisions on suitable locations and times of
exhibiting the artworks. For example, light-sensitive objects can be
safely exhibited nowadays in spaces where protection from natu-
ral radiation has been confirmed. In contrast, galleries illuminated
with a significant component of natural radiation are now intended
for works with moderate to low sensitivity. The data recorded in
these areas has indicated the need for restricting the exhibition
periods for objects with moderate light-sensitivity, with the aim of
minimizing their exposure to natural radiation.
5. Conclusions
A light levels assessment of several exhibition spaces of the
Donald W.Reynolds Center for American Art and Portraiture has
been conducted. The proposed methodology has allowed quan-
tifying and understanding the behavior of natural lighting inside
the building. The method has been validated by comparing win-
ter and summer data for various museum test-points. Simulated
and measured illuminance data shows good correlation at differ-
ent times of the year and under various sky illuminance conditions.
This methodological approach allows separating the radiation
into its natural and artificial lighting components. Cumulative
exposures were calculated by determining the area under the
illuminance curves using numerical integration. After evaluat-
ing the natural lighting data, four major zones of exposure were
identified. The annual natural lighting exposures obtained have
been correlated with maximum recommended values found in the
museumlighting literature.With theexceptionof exhibition spaces
located directly beneath skylights, it has been observed that the
majority of areas evaluated received total exposures which were
under the maximum allowable limits. Analysis of the cumulative
exposure data also permitted to identify two different types of
behavior, namely A and B, which were based on meter location
within a gallery and their orientation towards the sun. Exhibi-
tion areas exhibiting type A behavior experienced two exposure
maxima near the equinoxes while type B zones registered max-
imum exposures near the summer solstice. Small amounts of
UV radiation were detected at four south-wing locations. Even
though these UV radiation amounts remain under the 30Wlm−1
recommended level for very sensitive objects, preventive mea-
sures are now being implemented in the museum to further
reduce it.
This study has demonstrated the importance of conducting
long-term systematic studies of daylight illumination distribution
in museum buildings that make use of both electric and natural
lighting in their exhibition spaces. It is necessary to quantify the
sunlight since it shows a higher degree of variation throughout the
year and it is more difficult to control than artificial illumination.
Traditional museum light levels guidelines are based on correct
perceptionof theobjects andmake littleorno indicationabout their
preservation implications. Other guidelines only provide recom-
mended exposure limits that are basedonknown light-sensitivities
of various standard materials. The placement of an illuminance
meter next to each object is an unrealistic task and such an ambi-
tious program presents evident elevated costs for any museum.
However, once the distribution of natural lighting is known for a
specific location and orientation one will expect similar behaviors
in adjacent galleries with similar window sizes and floor areas. For
instance, natural lighting exposures and trends similar to the ones
described in this paper are expected for nearby museums in the
Washington DC area. Spot measurements taken near the surface
of objects at random times or days of the year only provide an
overall view of a particular lighting scenario at a specific moment.
These point measurements do not offer detailed information about
continuous changes in natural lighting exposure, which take place
throughout the year. Therefore, a qualitative and quantitative eval-
uation of natural lighting is essential in order to determine if there
Author's personal copy
64 J.M. del Hoyo-Meléndez et al. / Journal of Cultural Heritage 12 (2011) 54–64
is a significant contribution that may result in high risk of damage
for the artwork.
The proposed methodology can be applied to other cultural
institutions around the world. From a qualitative point of view,
daylight trends similar to the ones described in this paper are
expected in museums and galleries located in the Northern Hemi-
sphere. Based on the results, these annual daylight behaviors could
be easily identified since they are mainly influenced by location
and orientation of the area of interest with respect to various solar
positions observed throughout the year. If the results presented
in this study are used as a reference point, museums located in
lower latitudes would experience less variation during the year
due to sun paths, while institutions situated in the Southern Hemi-
sphere would expect similar cumulative exposure trends during
the equinoxes, accompanied by inverted natural radiation totals
during the solstices. In the present study, the illuminance data and
annual exposures recorded at various equivalent locations showed
good correlation, indicating that a lower number of monitors could
be employed when assessing natural radiation. This makes possi-
ble to extrapolate the results from representative galleries to other
similar spaces, since the distribution of daylight should be similar
for all equivalent areas. The results have also revealed that electric
lighting adjustments take place throughout the year. As a conse-
quence, variations in illuminance baselines or longer illumination
periods than anticipated may be observed due to special events
or gallery maintenance activities. Therefore, potential users must
carefully evaluate the electric lighting component of the examined
representative gallery since the amount of natural radiation is cal-
culated by subtracting the artificial lighting component from the
total exposure.
Acknowledgements
This research was supported by the Smithsonian’s Collections
Care Pool Fund, administered by the Smithsonian Collections Advi-
sory Committee and through a predoctoral fellowship from the
Smithsonian Institution Office of Research and Training Services.
The work was jointly sponsored by the Smithsonian American Art
Museum, the National Portrait Gallery and the Museum Conserva-
tion Institute. The authors would like to express their gratitude to
staff from both museums for all their support, valuable discussions
and recommendations concerning this work.
References
[1] G. Thomson, The museum environment, Butterworth-Heinemann, London,
1986.
[2] V. Daniel, Buildingmanagement andmaintenance in the newmillennium, Libr.
Rev. 50 (2001) 407–416.
[3] L. Bullock, Light as an agent of deterioration, in: National Trust (Ed.), The
National Trust manual of housekeeping: the care of collections in historic
houses open to the public, Butterworth-Heinemann, Oxford, 2006, pp. 92–101.
[4] M. Cassar, Environmental management: guidelines for museums and galleries,
Routledge, London, 1994.
[5] C. Cuttle, Light for art’s sake: lighting for artworks and museum displays,
Butterworth-Heinemann, Oxford, 2007.
[6] A. Derbyshire, J. Ashley-Smith, A proposed practical lighting policy for works of
art on paper at the Victoria and Albert Museum, in: D. Grattan (Ed.), Proceed-
ings of ICOM-CC, 12th Triennial Meeting, Lyon, 29 August – 3 September 1999,
International Council of Museums-Committee for Conservation, Paris, 1999,
pp. 38–41.
[7] K.M. Colby, A suggested exhibition policy for works of art on paper, J. Int. Instit.
Conserv. Canadian Group 17 (1992) 3–11.
[8] Illuminating andEngineering SocietyofNorthAmerica,Museumandart gallery
lighting: a recommended practice, Illuminating and Engineering Society of
North America, New York, 1996.
[9] S. Brown, I. Cole, V. Daniel, S. King, C. Pearson, Guidelines for environmental
control in cultural institutions, Heritage Collections Council, Canberra, 2002.
[10] J. Tétreault, Airborne pollutants in museums, galleries and archives: risk
assessment, control strategies, and preservation management, Canadian Con-
servation Institute, Ottawa, 2003.
[11] CIE 157:2004, Control of damage to museum objects by optical radiation, Com-
mission Internationale de L’Éclairage, Austria, 2004.
[12] B.L. Ford, Monitoring color change in textiles on display, Stud. Conserv. 37 (1)
(1992) 1–11.
[13] S. Michalski, Damage to museum objects by visible radiation (light) and ultra-
violet radiation (UV), in: Lighting in museums, galleries, and historic houses,
Papers of the conference, Bristol, 9–10th April 1987, TheMuseums Association,
London, 1987, pp. 3–16.
[14] R.L. Feller, Control of deteriorating effects of light upon museum objects,
Museum 17 (2) (1964) 57–98.
[15] L. Bullock, D. Saunders, Measurement of cumulative exposure using Blue Wool
standards, in:D. Grattan (Ed.), Proceedings of ICOM-CC, 12th TriennialMeeting,
Lyon, 29 August – 3 September 1999, International Council of Museums-
Committee for Conservation, Paris, 1999, pp. 21–26.
[16] M. Bacci, C. Cucci, A.L. Dupont, B. Lavédrine, M. Picollo, S. Porcinai, Dispos-
able indicators for monitoring lighting conditions in museums, Environ. Sci.
Technol. 37 (24) (2003) 5687–5694.
[17] F.G. France, Lighting the treasures, in: S. Thomassen-Krauss (Ed.), Facing imper-
manence: exploring preventive conservation for textiles, Proceedings of 6th
biennial North American Textile Conservation Conference, Washington DC,
November 6–9, 2007, North American Textile Conservation Conference, New
York, 2007, pp. 119–30.
[18] G. Martin, B. Pretzel, N. Umney, Preventive conservation in practice, V&A Con-
serv. J. 6 (1993) 15–18.
[19] G.M. Moeller, C. Weeks, AIA guide to the architecture of Washington, DC, Johns
Hopkins University, Baltimore, 2006.
[20] Tele Atlas 2008, EZ-locate program, Geocode, New Hampshire, 2008.
http://www.geocode.com/EZLI/ (accessed October 15, 2007).
[21] E.C. Boes, Fundamentals of solar radiation, in: J.F. Kreider, F. Kreith (Eds.), Solar
energy handbook, McGraw-Hill, New York, 1981.
[22] G.N. Tiwari, Solar energy: fundamentals, design, modelling and applications,
Alpha Science International, Pangbourne, 2002.
[23] J.J. Kim, Windows and daylighting group, Superlite 2.0, Lawrence Berkeley Lab-
oratory, Berkeley, 1993.
[24] J.C. McVeigh, Sun power: an introduction to the applications of solar energy,
Pergamon Press, New York, 1983.