The lunar soil spectra on this web site were measured for a paper submitted to 
the journal Icarus in March of 1997.  Below is an abbreviated summary of the 
paper without figures which should provide some context in which to understand 
the soil spectra.  The interested reader is referred to the journal for more 
information.


THERMAL INFRARED SPECTRA OF LUNAR SOILS

by

John W. Salisbury, Johns Hopkins University, Baltimore, MD

Abhijit Basu, Indiana University, Bloomington, IN

Erich M. Fischer, McKinsey & Co., Inc.


ABSTRACT


We have measured the infrared (2.08-14 µm) directional hemispherical 
reflectance spectra of lunar soils representing the major lithologic units so far 
sampled on the lunar surface, and soils of different exposure ages within those 
units.  Such reflectance (R) spectra can be used to calculate absolute emissivity 
(E) using Kirchhoff's Law (E=1-R).  The effects of exposure age vary with 
wavelength region.  In the 2-5 µm and 8-14 µm regions, lunar soils darken with 
exposure age, consistent with spectral behavior in the VNIR, and the dominant 
optical effect of increasing amounts of finely divided metallic iron in more mature 
soils.  However, in the 5-8 µm region soils tend to show higher reflectances with 
greater exposure age, which suggests some unanticipated change in the optical 
properties of fine metallic iron at those wavelengths.  The most useful spectral 
feature for compositional remote sensing is the Christiansen reflectance 
minimum (emissivity maximum), the spectral contrast of which is enhanced by 
the lunar environment, and the wavelength position of which is little affected by 
exposure age.  By contrast, the weak and relatively few overtone/combination 
tone absorption bands in the volume scattering region between 3 and 8 µm 
appear to be of limited usefulness.  The reststrahlen bands are also very weak in 
absolute emissivity spectra, and are evidently not enhanced by the lunar 
environment in the same fashion as the Christiansen feature.  Thus, they can 
only be used for remote sensing with measurements of extraordinarily high 
signal-to-noise (1000/1).  However, these features, as well as the transparency 
feature, do contain important mineralogical information, such as the relative 
abundances of plagioclase and pyroxene, and can be used for laboratory studies 
of lunar soils.  More certain and more quantitative mineralogical analyses of lunar 
soils appear feasible after additional spectral analysis of soil separates, and 
additional petrologic analysis of soil samples for which spectral data are 
available.



INTRODUCTION


Since the earliest days of spectroscopic remote sensing of the lunar surface, 
electronic transition bands exhibited by lunar soils in the visible and near-infrared 
(VNIR) regions of the spectrum have been used to determine mineralogical 
composition (e.g.,McCord and Johnson, 1970; McCord et al., 1972; Adams, 
1974).  There was also early recognition of the effects of space weathering on 
lunar soils and on their spectral reflectance (e.g., McKay et al., 1974; Hapke et 
al., 1975; Morris, 1976; Charette et al., 1976; and Basu, 1977.), resulting in a 
basic model which does, however, continue to undergo refinement (e.g., Pieters 
et al., 1993; Fischer and Pieters, 1994; and Clark and Johnson, 1996).  Briefly, 
lunar soils are derived principally by comminution of the underlying rocks by 
hypervelocity impact.  Over time, a rain of micrometeorite impacts continues the 
comminution process but, along with solar wind sputtering, also produces 
progressively more melt and vapor phase from such a particulate layer.  
Reduction of Fe2+ results in submicroscopic iron metal particles in the glass and 
deposited from vapor on the surfaces of grains.  Although other processes affect 
an aging soil, it is the iron particles in and on grains, primarily in glass-welded 
agglutinates, that principally account for the darkening and reddening of lunar soil 
reflectance with age of exposure on the lunar surface, as well as for the decrease 
in spectral contrast of the electronic transition bands.  Consequently, a good 
method for relating the exposure age of lunar soil to its changing optical 
properties is to use ferromagnetic resonance to measure the intensity of the 
characteristic resonance from single-domain iron within the soil normalized to the 
FeO content, yielding the widely used exposure index, Is/FeO (Morris, 1976).  
This approach was used by Fischer (1995) to select suites of lunar soils of similar 
composition but different exposure ages to study the effect of lunar space 
weathering on reflectance spectra, a model which we also follow in this paper.

In contrast with the range and sophistication of research done in the VNIR, much 
less is known about the spectral behavior of lunar soils in the thermal infrared.  
Early work by Lyon (1965) showed that the fundamental Si-O stretching vibration 
bands (the "reststrahlen bands") of silicates are greatly diminished in intensity, or 
spectral contrast, as particle size is reduced, making them difficult to detect in 
emission from a particulate surface such as the lunar regolith.  

Conel (1969) demonstrated that there exists an ancillary spectral feature, only 
indirectly related to the reststrahlen bands, that is relatively unaffected by particle 
size and can be used to determine composition.  More than one explanation has 
been offered for this so-called Christiansen feature (Hapke, 1996).  The 
traditional explanation for this feature, located near but not at the Christiansen 
frequency, is that it occurs in a wavelength region where the real part of the 
refractive index undergoes rapid change and thus may approach the refractive 
index of the medium (air or vacuum) surrounding the mineral grains, resulting in 
minimal scattering.  Because this takes place at a slightly shorter wavelength 
than the fundamental molecular vibration band, absorption is still relatively low.  
With little backscattering and little absorption, infrared radiation can pass through 
a sample relatively easily, resulting in a minimum in reflectance or a maximum in 
emittance.  This explanation agrees with experimental results, such as those 
obtained in vacuum described below, and is preferred by us.  Whatever its origin, 
the wavelength of this Christiansen feature was related to mineral composition by 
Conel (1969) and to igneous rock type by Logan et al. (1973), and Salisbury and 
Walter (1989).

Logan and Hunt (1970) found that a vacuum environment can introduce thermal 
gradients in a particulate sample that will significantly increase spectral contrast 
of the Christiansen emission peak.  In a vacuum, the lack of interstitial gas 
molecules results in very slow heat transfer between grains primarily by 
conduction across the small asperity points of contact, compared to the more 
rapid convective heat transport when gas is present.  For silicates, the optical 
depth in the visible region of the spectrum is greater than it is in the thermal 
infrared region.  Consequently, visible radiation penetrates to a greater depth in 
the sample than the depth from which particles are radiating heat into space.  
Logan and Hunt (1970) showed that under such conditions a cold skin of 
particles develops over a warmer interior.  Given such a thermal gradient, 
radiation passing relatively easily through the sample in the region of the 
Christiansen emission peak arises primarily from the relatively warm interior.  In 
the reststrahlen bands, however, where the absorption coefficient is high, 
radiation from the warm interior is absorbed before it reaches the surface and, 
consequently, the relatively cool surface layer dominates sample emittance at 
these wavelengths.  In effect, different temperatures are being observed at 
different wavelengths.  The net result, if the brightness temperature at the 
Christiansen peak is assumed to be the kinetic temperature (which is the usual 
assumption, because this is the wavelength of maximum emissivity), is a 
calculated emissivity curve that displays a greatly enhanced Christiansen peak 
compared to the reststrahlen band region, because of the disparity in radiance 
associated with the difference in effective temperatures.  It is not clear that this 
increase in overall contrast of the spectral curve is accompanied by increased 
contrast of the individual spectral features within the reststrahlen region.  The 
spectra obtained in a simulated environment by Logan and Hunt (1970) did not 
appear to have enhanced reststrahlen band fine structure, and they concluded 
that the prominent Christiansen emission peak was the spectral feature most 
likely to provide compositional information.  More recent vacuum chamber 
experiments by Henderson et al. (1996) confirm the enhancement of the 
Christiansen peak, and appear to show that reststrahlen features are actually 
slightly reduced in spectral contrast in the vacuum environment.

A second order effect of the enhanced spectral contrast of the Christiansen 
emission peak is a systematic shift of the peak to shorter wavelength due to the 
increased steepness of the falloff in emissivity toward the reststrahlen band on its 
long wavelength side (Salisbury and Walter, 1989; Henderson et al., 1996).  
Salisbury and Walter (1989) documented this shift of the Christiansen feature by 
comparing spectral emission measurements of different powdered igneous rock 
types in vacuum (Logan et al., 1973) with their own directional hemispherical 
reflectance measurements of the same rock powders in air.  They showed that 
the shift was systematic and determined a correction factor.  They also showed 
that the wavelength of the Christiansen feature could be correlated with a 
chemical index that is a good indicator of igneous rock type. This SCFM index is 
the ratio of the oxide abundance of silica to the combined oxide abundances of 
silica, calcium, iron and magnesium.

Another spectral feature that may be useful for compositional remote sensing of 
particulate surfaces occurs in a wavelength region of relative transparency 
between regions of high absorption coefficient associated with the Si-O stretching 
and O-Si-O bending vibration bands.  These fundamental vibration bands are 
dominated by surface scattering, which is why the reststrahlen bands are 
reflectance peaks.  In between these bands, the absorption coefficient is 
sufficiently low that the silicate grains become optically thin and volume 
scattering comes to dominate the scattering process as particle size is reduced.  
This enhanced volume scattering also results in a broad reflectance peak, the 
wavelength of which has been shown to be related to mineral composition and 
terrestrial igneous rock type (Salisbury and Walter, 1989).

All of these measurements of terrestrial surrogates have been accompanied by 
relatively few measurements of the Moon and of returned lunar soils.  Remote 
sensing measurements of the Moon have been badly hampered by atmospheric 
absorption and other problems.  The controversial results have been reviewed by 
Nash et al. (1993) and Salisbury et al. (1995), and will not be considered here.  
Laboratory spectral emission measurements of returned lunar soils in a vacuum 
environment have shown qualitatively, however, the prominent Christiansen peak 
and barely discernible reststrahlen features expected from surrogate 
measurements (Logan et al., 1972; Salisbury et al., 1973).  Some of these 
laboratory spectra are reproduced in Figure 1a.
                                                                
	Figure 1a.  Emission spectra of Apollo 16 soil samples measured in a 
	simulated lunar environment.  Emissivity has been calculated assuming an 
	emissivity of 1.0 at the Christiansen peak, and spectra are displaced 
	vertically by 0.1 increments (from Salisbury et al., 1973). 

To date, only one lunar soil has been measured in both emission in a simulated 
lunar environment and in directional hemispherical reflectance from which 
absolute emissivity can be calculated (Figure 1b).  Unfortunately, the emission 
spectrum does not span the entire wavelength region available for remote 
sensing through the atmospheric window, and has relatively low (1.5%) spectral 
resolution.  Still, it does show quantitatively the enhancement of the Christiansen 
peak in a vacuum environment and, despite the low resolution of emission 
spectrum, what appears to be reduced spectral contrast of the reststrahlen 
features.  However, this soil sample had the lowest spectral contrast of all of the 
soil samples measured by Logan et al. (1972) and Salisbury et al. (1973) in a 
vacuum environment, making the difference so small between the two 
measurements that no firm conclusion can be reached.

                                                                
	Figure 1b.  Comparison of spectra of lunar soil 10084 measured in 
	emission in a simulated lunar environment and in directional hemispherical 
	reflectance in air (emission measurement from Logan et al., 1972).  Note 
	that the shift of wavelength associated with the thermal gradient imposed 
	by the simulated lunar environment is very small for such mafic rocks, as 
	found by Salisbury and Walter (1989).  The expanded scale shows subtle 
	reststrahlen features.


Other laboratory measurements of lunar soil spectra in the thermal infrared are 
few in number, and are difficult to relate to the measurements of lunar soils and 
soil surrogates outlined above for a variety of reasons.  Aronson and Smith 
(1978) and Aronson et al. (1979) made some spectral measurements in a 
vacuum environment, but presented their data in a brightness temperature format 
instead of calculated emissivity.  Still, they do show brightness temperature 
curves for samples 67711 and 10084, and conclude that the former has the most 
spectral contrast of any soil sample they measured, and the latter has the least, 
which is consistent with the measurements shown in Figure 1.  It is also 
consistent with the findings in this paper about the effects of soil exposure age.

Nash (1991) made biconical reflectance measurements, which cannot be 
quantitatively converted to emissivity (Nicodemus, 1965).  In addition, the 
samples were slightly packed, which slightly enhances the reststrahlen features.  
Again, however, there is good qualitative agreement with conclusions drawn 
above.  That is, spectra of rock chips display greater spectral contrast than the 
spectra of disaggregated breccias, and much greater contrast than the spectra of 
soils.  

It is the purpose of this paper to: 1) provide the first directional hemispherical 
reflectance spectra (interpretable in terms of absolute emissivity) of soil samples 
representative of all of the major lunar lithologic units sampled so far; 2) 
document the effect of exposure age on the thermal infrared spectral properties 
of these soils; 3) determine the extent to which the wavelength of the 
Christiansen peak can be correlated with the SCFM chemical index of lunar soils, 
and whether or not soil exposure age affects this correlation; 4) identify 
reststrahlen features in lunar soil spectra that are due to different minerals; and 
5) determine the extent to which mineral abundances estimated from reststrahlen 
features correlate with the normative mineral abundances calculated from soil 
chemistry.

We recognize the need for caution in applying our findings directly to remote 
sensing of lunar composition because of the effect of the lunar environment on 
the shape of the spectral curve, especially the magnitude and wavelength of the 
Christiansen peak.  However, we believe that it is best to start with absolute 
emissivity and then account for lunar environmental effects, as has been done for 
the wavelength position of the Christiansen peak by Salisbury and Walter (1989), 
rather than to use qualitative data obtained with an always questionable 
environmental simulation.  Further, the results reported here can be directly 
applied to laboratory investigations of lunar soils.


SAMPLE SELECTION

In the absence of any accurate means to artificially mimic the evolution of lunar 
soils, Fischer (1995) selected suites of returned soils of similar compositions but 
different exposure ages, which  were then assumed to approximate the evolution 
of a soil of a given composition.  To this end, he identified soils with a wide range 
of exposure age as quantified by Is/FeO (Morris, 1976), which were within 1 wt.% 
FeO, Al2O3, and TiO2 of selected immature soils.  Soils selected in this manner 
which displayed VNIR absorption band centers different from their counterparts 
(indicative of mineralogical differences) were discarded.  This process resulted in 
selection of four compositionally distinct suites of three to four soils each, with 
each soil sample in a suite presumably representing a distinct evolutionary stage 
of soil development.  We have adopted this model and used the fine fraction (< 1 
mm) of each of these soils as representative of the surface layer.  In addition, we 
have selected an additional mare soil (10084) to extend the soil types to include 
a high titanium mare soil, and have added several other soil samples similar to 
those originally selected by Fischer (1995) to guard against possible bias 
introduced by too few samples.  The Fischer sample suites are described below 
and all sample selections are documented in Table 1.

Soil Suites

The Apollo 16a Reference Suite:  This suite, consisted of four soils: 61221, 
67701, 60051 and 64801, in order of increasing exposure age as determined by 
the Is/FeO index (see Table 1).  Whereas most Apollo 16 soils can be interpreted 
as mixtures of at least two or three lithologic components (Houck, 1982; Basu 
and McKay, 1984), the group of soils represented by this reference suite is 
characterized by relatively higher aluminum and lower iron and titanium 
concentrations than other Apollo 16 soils, and is likely derived dominantly, but not 
purely, from the friable light matrix breccia unit sample most notably by the North 
Ray crater impact.

The Apollo 16b Reference Suite:  This suite consists of soils 67941, 61241, and 
68501, in order of increasing exposure age as determined by the Is/FeO index 
(see Table 1).  This suite, slightly less aluminous than the Apollo 16a suite, was 
selected by Fischer (1995) to determine whether very subtle compositional 
differences relative to the Apollo 16a suite might have noticeable effects upon 
optical alteration.  As Fischer (1995) found in the VNIR, we find that the subtle 
compositional differences are reflected in significant differences in the thermal 
infrared spectra as described below in the Results section.

The Apollo 14 Reference Suite:  This suite consists of soils 14141, 14149, and 
14148, in order of increasing exposure age as determined by the Is/FeO index 
(see Table 1).  This suite represents materials transitional between highlands and 
mare in terms of major element chemistry.

The Apollo 12 Reference Suite:  The Apollo 12 suite consists of soils 12030, 
12024, and 12023, in order of increasing exposure age as determined by the 
Is/FeO index (see Table 1).  This suite is the one mare reference suite for 
analysis of exposure age effects.  It represents a low-titanium mare lithology.


INSTRUMENTATION AND MEASUREMENT TECHNIQUE

Laboratory Directional Hemispherical Reflectance Measurements

Directional hemispherical reflectance measurements were made at 4-
wavenumber resolution using a Nicolet FTIR spectrometer with a gold-coated 
integrating sphere and a liquid nitrogen-cooled, mercury-cadmium-telluride 
detector.  Samples were measured at a wall-mounted port at the bottom of the 
sphere, and illuminated at 10¡ off of the vertical through a port in the top of the 
sphere.  Spectra were calibrated to absolute reflectance using a gold mirror and 
standards traceable to NIST (Salisbury and Walter, 1989).

As shown by Nicodemus (1965), such 10¡ directional hemispherical reflectance 
measurements (R) can be used to predict directional spectral emissivity (E) using 
Kirchhoff's Law (E=1-R).

Sample Preparation

Each lunar soil sample (<1 mm fraction) was poured onto a fresh, waxed 
weighing paper from its container as received from the Lunar Sample Curator, 
and the paper was gently shaken sideways to flatten the sample prior to spectral 
measurement while still on the weighing paper.

A better simulation of the underdense nature of the upper surface of the lunar 
regolith might have been obtained by sifting the samples, but we did not want to 
risk the sample contamination that would inevitably result.  One sample was 
measured both sifted through a 75 µm screen and poured as described above to 
illustrate the spectral differences that might result from the smaller grain size and 
higher porosity at the surface of the sifted sample.  As shown in Figure 2, the 
sifted sample has a considerably higher reflectance in the 2-8 µm region, and a 
slightly lower reflectance in the 8-14 µm region.  This spectral behavior is the 
result of the changing opacity of silicate minerals with wavelength.  That is, in the 
wavelength region where the sample is relatively transparent there is an increase 
of reflectance with decreased particle size, because more photons survive 
passage through the smaller gains to be backscattered to the observer.  In 
regions of high opacity, on the other hand, there is a net loss of photons due to 
the absorption that occurs with the increased volume scattering associated with 
smaller grain size and higher porosity (Salisbury and Wald, 1992).  The change 
in spectral reflectance between sifted and poured samples is relatively small in 
the reststrahlen bands, where the sample is most opaque, because the 
interaction of radiation with the sample occurs at or near the sample surface.  
Thus, even when more large grains are present at the surface, as in the case of 
the poured sample, the smaller particles with which they are invariably coated  
tend to dominate this near-surface interaction (Salisbury and Wald, 1992).  


REFERENCES


Adams, J. B. 1974. Visible and near-infrared diffuse reflectance spectra of 
pyroxenes as applied to remote sensing of solid objects in the solar system. J. 
Geophys. Res. 79, 4829-4836.

Aronson, J. R. and E. M. Smith 1978. Mid infrared spectra of lunar and analog 
soils. Proc. Lunar Sci. Conf. 9th, 2911-2917.

Aronson, J. R., A. G. Emslie, E. M. Smith, and P. F. Strong 1979. Infrared 
spectra of lunar soils and related optical constants. Proc. Lunar Sci. Conf. 10th, 
1787-1795.

Basu, S. 1977. Steady state, exposure age and growth of agglutinates in lunar 
soil. Proc. Lunar Sci. Conf. 8th, 3617-3632.

Basu, A. and D. S. McKay 1984. Petrologic comparisons of Cayley and 
Descartes on the basis of Apollo 16 soils from stations 4 and 11. Proc. Lunar Sci. 
Conf. 14th, B535-B541.

Charette, M. P., L. A. Soderblom, J. B. Adams, and T. B. McCord 1976. Age-
color relationships in the lunar highlands. Proc. Lunar Sci. Conf. 7th, 2579-2592.

Clark, B. E. and R. E. Johnson 1996. Interplanetary weathering: Surface erosion 
in outer space. EOS. 77, 141-145.

Conel, J. E. 1969. Infrared emissivities of silicates: Experimental results and a 
cloudy atmosphere model of spectral emission from condensed particulate 
mediums. J. Geophys. Res. 74, 1614-1634.

Estep, P. A., J. Kovach, and C. Karr 1971. Infrared vibrational spectroscopic 
studies of minerals from Apollo 11 and 12 lunar fines. Lunar Planet. Sci. Conf. 
2nd, 2134-2151.

Fischer, E. M. 1995. Quantitative compositional analysis of the lunar surface from 
reflectance spectroscopy: Iron, aluminum, and a model for removing the optical 
effects of space weathering. Ph.D. thesis, Brown University, 194 pp.

Fischer, E. M. and C. M. Pieters 1994. Remote determination of exposure degree 
and iron concentration of lunar soils using VIS-NIR spectroscopic methods. 
Icarus. 111, 475-488.

Hapke, B., 1996. Applications of an energy transfer model to three problems in 
planetary regoliths: The solid-state greenhouse, thermal beaming, and emittance 
spectra. J. Geophys. Res. 101,16,833-16,840.

Hapke, B., W. Cassidy, and E. Wells 1975. Effects of vapor-phase deposition 
processes on the optical, chemical, and magnetic properties of the lunar regolith. 
The Moon  13, 339-353.

Henderson, B. G., P. G. Lucey, and B. M. Jakosky 1996. New laboratory 
measurements of mid-IR emission spectra of simulated planetary surfaces. J. 
Geophys. Res. 101, 14,969-14,975.

Houck, K. J. 1982. Petrologic variations in Apollo 16 surface soils Proc. Lunar 
Sci. Conf.13th, A197-A209.

Linsay, J. F. 1976. Lunar Stratigraphy and Sedimentology. Elsevier Scientific 
Publishing Co., New York, 302 pp.

Logan, L. M., and G. R. Hunt 1970. Emission spectra of particulate silicates 
under simulated lunar conditions. J. Geophys. Res. 75,  6539-6548.

Logan, L. M., G. R. Hunt, S. R. Balsamo, and J. W. Salisbury 1972. Midinfrared 
emission spectra of Apollo 14 and 15 soils and remote compositional mapping of 
the moon. Proc. Lunar Sci. Conf. 3rd.,  3069-3076.

Logan, L. M., G. R. Hunt, J. W. Salisbury, and S. R. Balsamo 1973. 
Compositional implications of Christiansen frequency maxima for infrared remote 
sensing applications. J. Geophys. Res. 78, 4983-5003.

Lyon, R. J. P. 1965. Analysis of rocks by spectral infrared emission (8-25 
microns). Econ. Geol. 60, 717-736.

McCord, T. B. and T. V. Johnson 1970. Lunar spectral reflectivity(0.30 to 2.50 
microns) and implications fro remote mineralogical analysis. Science. 169, 855-
858.

McCord, T. B., M. P. Charette, T. V. Johnson, L. A. Lebofsky, C. Pieters, and J. 
B. Adams 1972. Lunar spectral types. J. Geophys. Res. 77, 1349-1359.

McKay, D. S., R. M. Fruland, and G. H. Heiken 1974. Grain size and the 
evolution of lunar soils. Proc. Lunar Sci. Conf. 5th, 887-906.

Morris, R. V. 1976. Surface exposure indices of lunar rocks: A comparative FMR 
study. Proc. Lunar Sci. Conf. 7th, 315-335.

Murcray, F. H., D. Murcray, and W. Williams 1970. Infrared emissivity of lunar 
surface features. I. Balloon-borne observations. J. Geophys. Res. 75, 2662-2669.

Mustard, J. F., and J. E. Hays 1997. Effects of hyperfine particles on reflectance 
spectra from 0.3-25 µm. Icarus.  125, (in press).

Nash, D. B. 1991. Infrared reflectance spectra (4-12 µm) of typical lunar samples. 
Geophys. Res. Lett. 18, 2145-2147.

Nash, D. B., and J. W. Salisbury 1991. Infrared reflectance spectra of plagioclase 
feldspars. Geophys. Res. Lett . 18, 1151-1154.

Nash, D. B., J. W. Salisbury, J. E. Conel, P. G. Lucey, and P. R. Christensen 
1993. Evaluation of infrared emission spectroscopy for mapping the Moon's 
surface composition from lunar orbit. J. of Geophys. Res. 98, 23,535-23,552.

Nicodemus, F. E. 1965. Directional reflectance and emissivity of an opaque 
surface. Appl.. Opt. 4, 767-773.

Pieters, C. M., E. M. Fischer, O. Rode, and A. Basu 1993. Optical effects of 
space weathering: The role of the finest fraction. J. Geophys. Res. 98, 20,817-
20,824.

Salisbury, J. W. and A. Wald 1992. The role of volume scattering in reducing 
spectral contrast of reststrahlen bands in spectra of powdered minerals. Icarus. 
96, 121-128. 

Salisbury, J. W., and L. S. Walter 1989. Thermal infrared (2.5-13.5 µm) 
spectroscopic remote sensing of igneous rock types on particulate planetary 
surfaces. J. Geophys. Res. 94, 9192-9202. 

Salisbury, J. W., G. R. Hunt, and L. M. Logan 1973. Infrared spectra of Apollo 16 
fines. Lunar Planet. Sci. IV 3, 3191-3196.

Salisbury, J. W., L. S. Walter, Norma Vergo, and D. M. D'Aria 1992. Mid-Infrared 
(2.1 to 25 mm) Spectra of Minerals, 267 pp, Johns Hopkins University Press, 
Baltimore.

Salisbury, J. W., D. G. Murcray, W. J. Williams, and R. D. Blatherwick 1995. 
Thermal infrared spectra of the Moon. Icarus. 115, 181-190.

Thomson, J. L., and J. W. Salisbury 1993. The mid-infrared reflectance of mineral 
mixtures (7-14 µm). Remote Sens. of Environ. 45, 1-13.

Walter, L. S., and J. W. Salisbury 1989. Spectral characterization of igneous 
rocks in the 8- to 12-µm region. J. Geophys. Res. 94, 9203-9213.