The Rainband Spectroscope
Peter Abrahams,   telscope@europa.com

Illustrations:
The solar spectrum (top); spectrum with a rainband that is darker & more 
distinct than usual (bottom):
     http://home.europa.com/~telscope/spctrscp-rnbnd.jpg    20 kb
The rainband seen at various times, illustration from Schellen, Spectrum 
Analysis, 1872.
     http://home.europa.com/~telscope/spctrm-Schellen.jpg    90 kb
A pocket rainband spectroscope:
     http://home.europa.com/~telscope/spct-rnbnd.jpg    54 kb

   Hand held 'direct vision' spectroscopes were made by several British, 
French, and German instrument makers, and commonly sold in the late 1800s, 
though no doubt the price kept them from truly widespread use.  The popular 
interest had been captured by the development of the science of spectroscopy, 
especially by astronomical studies of the composition of stars, which for so 
long had been thought to be beyond our knowledge.  A direct vision spectroscope 
is simply a model that offers dispersion with no deviation, so that it can be 
pointed directly at a light source, which makes it considerably easier to use 
as a hand held instrument.  The instrument consists of a small tube that is 
pointed at the light, which passes through a slit, a collimating lens, the 
prism train (usually 3 crown glass prisms and 2 flint glass prisms cemented 
together), and exits through a magnifying lens or a plain glass cover.  When 
viewing sunlight through a spectroscope, the fine black Fraunhofer lines are 
sometimes accompanied by other features.  When the air is carrying a large 
quantity of water vapor over a long distance along the line of sight, a red 
band can be observed adjacent and to the red side of the D lines of sodium; the 
band varying in darkness with the concentration of water vapor.  The 
spectroscope shows whether there is more or less than the usual quantity of 
water vapor in the otherwise dry gases in the upper parts of the atmosphere.  
The rainband shows warm moist air at high elevations, which usually is rising 
to produce rain at some location.  If this water vapor is observed in the 
direction from which the wind is blowing, it often means that rain is 
forthcoming.  This phenomenon was developed in the late 1800s into a form of 
weather forecasting.

   The direct vision prism spectroscope was invented in 1860 by Giovanni Amici, 
but was derived from the ideas of John Dollond & others.  The pocket direct 
vision spectroscope was patented in 1861 by Browning and Crookes (British 
Patent 1181).    (Source: Austin)
   Absorption lines from earth's atmosphere were known since Bunsen & 
Kirchhoff, circa 1862.  During the 1860s, Josiah Cooke at Harvard described 
atmospheric lines from water vapor.  Astronomical spectroscopists learned that 
weather conditions had an effect on their measurements.    (Peterson)
   Jules Janssen in 1864 built a 5  prism spectroscope that resolved the 
atmospheric bands into very fine lines and more clearly showed their varying 
intensity, darkest at dawn and dusk.  He traveled to the peak of 3000 meter 
Faulhorn to observe upper elevation spectra, finding that the bands were much 
fainter than observed at lower altitudes, and taking this as further evidence 
of terrestrial origin.  Later, in October 1864, he arranged for a large stack 
of pine wood to be burned near Geneva, and observed the spectra from a distance 
of 21,000 meters and from a short distance, noting that some of the lines & 
bands were absent when the fire was observed from nearby.  Secchi also 
performed this experiment, viewing the spectrum of a flame from 2,000 meters, 
and causing large bonfires to be lit on a mountain for spectral observations.    
(Schellen)
   In 1866, Janssen obtained from the Paris Gas Company an iron pipe 118 feet 
in length, closed at both ends with plate glass, which he filled with steam 
under seven atmospheres of pressure.  Flame from 16 gas burners outside one 
window was viewed with the spectroscope from the other window.  In the area 
between the red region and the D line, groups of dark lines were seen, 
resembling those in the spectrum of the setting sun.  Janssen also demonstrated 
absorption lines from water vapor in the infra-red area of the spectrum, and 
showed in the ultra violet a lesser darkening from moisture over the entire UV 
area of the spectrum.  Performing variations of this experiment, Janssen viewed 
steam under pressure & saw no effect with a 40 foot steam tube, slight effect 
with an 84 foot tube, and fully detailed effect with a 330 foot tube.  Using a 
telescope, Janssen observed the terrestrial water bands in the spectrum of 
Sirius.  He also observed, or believed he observed, water vapor in the 
atmospheres of Mars and Saturn.    (Schellen, Smyth 1877)
   Also in the 1860s, Henry Roscoe published work that noted observations of 
lines seen at sunset, as described in Angstrom's 'Normal Solar Spectrum'.   
(Smyth 1877)  Angstrom observed and wrote detailed descriptions of the water 
bands, published with elaborate drawings; noting that spectral bands which 
change with temperature are likely to be water, while bands that do not change 
are likely to be gases.   (Schellen)
   In France, M.A. Cornu separated the terrestrial lines from the solar lines 
by inferring that the rotation of the sun would cause its lines to shift, to 
the blue and then the red, as the sun rotates.  By observing the sun and 
alternating the view between the two solar hemispheres, the solar lines will 
move back & forth within the field.   (Schellen)
   Charles Piazzi Smyth, Astronomer Royal for Scotland 1845-1888, was the 
leading advocate of the spectroscope as a meteorological tool.  He had noticed 
the rainbands in his daily observation of the sky during his travels, and over 
time developed a correlation between the appearance of the bands and upcoming 
rain.  His first report was in 1875, when he wrote that his initial observation 
was in Palermo, in 1872.  Smyth visited Janssen in 1876 during a trip to Paris, 
and observed spectra of steam in Janssen's experiment.    (Austin, Peterson)
   H. Schellen in Germany wrote an authoritative text on spectroscopy and 
practiced the science as well.  Schellen used a direct vision spectroscope by 
Reinfelder & Hertel that showed the D line to be double.  He found that a 
marked development of the rainband was usually followed by rain, in the summer 
more so than the winter, even when skies were cloudless; and that the rain band 
could readily be seen throughout a 'dull day', but was well defined during a 
heavy rain and even darker at sunset.  Much practice was required to become 
proficient in recognizing darkening of the band.
   In England, the photographer William Abney, along with Mr. Festing, 
photographed spectral bands of water in the infra-red, circa the 1880s.  They 
found good correlation with rainfall; and made comparisons with light 
transmitted through varying distances of water.
   The term 'Rain band' was introduced in 1881 by J. Rand Capron, Fellow of the 
Royal Astronomical Society & Fellow of the Royal Meteorological Society.  
Capron noted that Angstrom's solar line diagrams showed variability with 
varying moisture in the atmosphere.  He made further observations using a 
McClean's star spectroscope with an added slit; believing that an adjustable 
slit & achromatic lenses were useful but not necessary.  In 1885, Capron 
suggested a comparison spectrum using light transmitted through a wedge prism 
of glass with trace didymium content, which shows a band similar to the 
rainband.  Both the width and the tint of the band were to be compared.
   John  Browning manufactured a spectroscope with 5 prisms, similar to other 
pocket spectroscopes but engraved 'Rainband Spectroscope'; and included 
information on their use in the second edition of his 'How to Work with the 
Spectroscope', with a color drawing to help.  Browning sold Capron's 1881 
booklet, 'A Plea for the Rain Band'.  Browning's rainband spectroscope was an 
attempt to standardize the instrument, since variations in slit width and 
dispersive qualities of the prisms would make rainband comparisons difficult.  
Capron and Browning worked together to promote the new instrument.  (Peterson, 
Austin)

   Any pocket direct vision spectroscope could be sold as a rainband 
instrument.  The 'Grace's Spectroscope' was a DVS that was sold especially for 
that purpose, with a larger aperture, a knurled wheel for focus adjustment, and 
an eye lens instead of plate glass.  It displayed a larger spectrum, showing 
lines & bands more clearly, and resolving rain bands into separate lines.  It 
was also usable closer to dawn & dusk.  (Peterson)
   John Browning's catalog of March 1923, 'Prismatic Instruments', included a 
section: 'Direct Vision Spectroscopes, Medium Size, Mechanical Focusing 
Adjustment'.  'This instrument was originally designed for meteorological use 
and has for many years been known as Grace's Rainband Spectroscope, but is 
equally suitable for all the various purposes for which Direct Vision 
Spectroscopes can be used. It has greater dispersion than the miniature type 
and the mechanical focusing adjustment will be found useful for faint lines and 
bands as distinct from the well-marked absorption bands, for which the simpler 
patterns of the Miniature type suffice.'  The text continues with detailed 
instructions for use of the above instrument as a Rainband Spectroscope.  
(Robert Ariail)
   Frederick Cory, Fellow of the Royal Meteorological Society, had Browning 
attach a comparison prism to his Grace's spectroscope, along with a micrometer 
scale, allowing comparison with a reference source.
   Adam Hilger made, promoted, and sold spectroscopes for meteorological use.
   Louis Casella also made & sold pocket spectroscopes, with Capron's booklet.
   Imports to the U.S. followed sales in England.  Winslow Upton of the U.S. 
Signal Service wrote on the subject of rainband spectroscopes in 1883, giving 
detailed instructions, but not regarding the instrument as entirely useful, 
since the band was minute and only a skilled, experienced user could make the 
observations.  Water vapor could be detected, but the conditions conducive to 
formation of rain from that vapor could not be observed.  A larger instrument 
was helpful.  He suggested comparisons with clear areas of sky, and maintaining 
proper focus and slit width.   (Peterson)
   Louis Bell, of the Physics department at Johns Hopkins, noted the misuse of 
the spectroscope by the 'weatherwise'.  The problem he saw was the 
interpretation of the dark quality of the rain band, by different observers 
under varying conditions and using assorted instruments.  Bell devised a simple 
photometer to mount on his 5 prism spectroscope made by Schmidt of Berlin.  A 
thin selenite window was placed in front of the slit, which produced over half 
the field a set of interference bands, each the approximate width of the widest 
rain band.  In the cap of the spectroscope was mounted a Nicol prism that 
rotated to darken the illuminated, clear area of the field to match the shade 
of the rainband.  A scale on the rim allowed calibration, and the tool allowed 
Bell to conclude that a proper rainband spectroscope, properly used, could be 
quite successful at predicting rain.   (Peterson)
   The rainband spectroscope gave subjective results, and was not adopted as 
the standard instrument hoped by its proponents, despite which they were sold 
into the 1920s.
   In the 21st century, the study of spectral water has moved to the stars.  
Absorption lines are found in the infrared spectra of cool stars including 
Betelgeuse & Antares, red supergiants with extended atmospheres.  The water 
lines originate in the atmosphere of the star, not in the circumstellar 
material; and in fact water is abundant in cool stars and is also found in 
sunspots.  In the study of variable stars, the spectrum of water changes with 
the luminosity of the star.   (Jennings)

   Various observers, representing different levels of skill in observing & 
describing, published their notes on rainband observations, some of which 
follow:
   The type of spectroscope used influences the effect that can be seen.  
Spectroscopes with higher magnification or increased dispersion often give an 
image that is too dark for rainband work.  Under very bright conditions, a 
larger spectroscope with greater dispersion, allows the rainband to fill the 
field of view, and shows several water lines & bands near and in between the D 
lines of sodium.
   Writers described the definitive broad band adjacent to the red side of the 
sodium line D1.  Progressing from D1 towards the far red, next are a pair of 
lines almost as widely spaced as the D lines and sometimes as well defined, 
next are two or three faint bands, next are three lines, the first of which can 
be blacker than either D line, next are three equally spaced bands, and then 
into the red area is a series of faint 'hazebands' -- all of which together 
make up the single rainband of a pocket instrument.  There are also several 
other less prominent rain bands in the spectrum, the second most visible band 
is found on the yellow side of the C line, and is often present when viewing 
through a mist, when the D band is usually absent.  Much variation in 
appearance is found; the bands can be narrow or broad, pale or dark, the edges 
can be abrupt or gradually fading, and a single band can be split into two half 
bands.  Rainbands are darker shades of the spectral coloration, they can be 
prominent but are usually subtle.
   When the solar Fraunhofer lines are sharp, the rainband is usually absent; 
and when solar lines are faint, the rainband is usually present.
   Schellen notes that the rain band can be readily seen throughout a 'dull 
day', but is particularly well defined during a heavy rain, and even darker at 
sunset.
   When the band is viewed at sunset, against a red sky, it is more distinct 
and darker; but when viewed against a yellow sunset sky, it is more difficult 
to see.
   Comparing the spectrum at sea level versus at a high elevation, the solar 
lines are unchanged with elevation, but terrestrial lines show differences.
   When it is actually raining, the band usually faint, except in summer, when 
it can be strong while it is raining.
   The rainband is much more noticeable in a warm climate, and harder to see in 
a cold climate.  The band for water vapor is commonly visible in summer, when 
temperature is warm and air is not dry.  Snow is not predicted in the rainband, 
the band is faint in cold air, when vapor is transformed into snow crystals.  
Temperature has an effect, but not on a daily cycle; as the air warms during 
the day, it increases its capacity to hold water but not the amount of water in 
the air.
   When, at a given temperature, it is viewed to be twice as dark as the 
commonly observed level of contrast, it may be regarded as an indicator of rain 
-- the atmosphere cannot hold this amount of moisture.
   An appearance of the rainband almost always means rain when wind conditions 
are right; but often there will be rain with no rainband.  
   Often the band is present in one area of the sky only.  If the band heavy at 
higher angles of view or at the zenith, it is more certain that there will soon 
be rain.  Rain can fall in very localized small areas after a rain band 
appears, and thus it can seem to have failed as a prognosticator, when in fact 
rain has fallen - but before the cloud reaches the observer or in a nearby 
location.
   The rainband can sometimes be seen in a fog or mist, when it does not 
indicate rain.
   Cory observed particular changes in the spectrum before and during 
thunderstorms.  The blue region became deeper and the green more prominent.
   If you cannot see the D line, hold a match in front of the slit as it 
ignites, and a bright yellow line is seen adjacent to where the rainband should 
be.
   Record keeping should include data on the quality of rainband for 8 points 
of the compass and for the zenith, and the direction of wind.

   Instructions for use, consolidated from various sources:
   Adjust the spectroscope.   First, close the slit until the jaws almost touch 
or until the Fraunhofer lines are sharp.  If the jaws are closed too tight, the 
spectrum grows dim and black lines sometimes appear, running lengthwise along 
the spectrum.  If these horizontal lines appear, they are usually from dust in 
the slit; for which the slit should be opened widely & a wood sliver used to 
wipe the slit edges -- using paper or cloth, or blowing on it, usually makes 
the dust worse.
   Observe in the direction from which the wind is blowing, or look at upper 
elevation clouds & view towards the direction they are coming from.  View at an 
angle of about 10 to 20 degrees above the horizon; a very long path through the 
atmosphere is required, but viewing at too low an angle will show a 'false 
band' due to moisture from the earth.  Observe a bright area of the sky, and 
compare the spectrum of the sky directly overhead, to the spectrum seen at a 
low angle, looking for a change in the area between the yellow and the red.  
The rainband is a darker shading, immediately adjacent to the D lines of 
sodium, on the red side, in a band about one quarter as wide as the yellow 
area.  It appears narrower and weaker with less moisture.  The width and 
darkness of the rainband increase with increasing moisture.  It is important to 
'scan vertically', check near the horizon, and then upward to 15-20 degrees, if 
strong band appears in both places, it is usually prognostic of rain, and a 
band that is strong to the zenith is very rare but indicates heavy rain.
   Comparison is helpful in detecting the rainband.  View the upper part of the 
sky & adjust the slit until you can see the stronger dark solar lines, D, E, b, 
and F.  View to the lower and brighter parts of the sky (but the brightness 
must be from directly transmitted light, not scattered light).  Observe any 
changes near the D line.  A darker streak on the red side of D, appearing to be 
partly attached to the D line, can be estimated to be of a certain darkness & 
compared over time.
   It helps to shade the eyes by cupping a hand around the observing eye; and 
it is important to focus the instrument so that the Fraunhofer lines are sharp; 
generally one needs to push in the focuser to focus on lines in the blue area & 
pull out for lines in the red area.  When the band is faint, use averted 
vision, and look directly at yellow area to try to see the rainband. 
   The sky must be well illuminated, since brightness provides adequate 
contrast in the spectrum.  Overcast skies with low clouds do not provide a long 
enough light path, and clouds can reduce contrast.  Smoke and dust prevent 
observations, by limiting the light path and by scattering light.  However, 
when the sky is very deep & dark blue, it is harder to see band
   Smyth (1877) estimated intensity by comparing the rainband with the 'low sun 
band' on the green side of D, which is seen through dry air only, at a maximum 
during sunrise and sunset, and visible throughout the day during winter at 
higher latitudes.  It is separated from the D line by a bright yellow band and 
can sometimes be seen very strongly just after sunset.

   There are inconsistencies in the contemporary explanations for the observed 
phenomena.   This is partly because the effects were not understood, but also 
because atmospheric water can be gaseous vapor, or water drops varying in size 
from very fine droplets of mist to raindrop size.
   The various texts describe rainbands as produced by water vapor only, or 
produced by liquid water in the air - as mist or raindrops, or produced by both 
vapor & liquid.  Others note that water vapor is revealed in a fine line 
spectrum, that might show as a band in a low resolution spectroscope; and that 
reflected light from visible droplets can cause very faint effects on spectrum 
(it is unclear whether they mean light reflected from the surface of drops, or 
light refracted through the drops).  Some observers felt that all rainbands, 
normally seen as a dark band or shading, can be shown to be a set of fine lines 
when using a larger spectroscope or under magnification.

   The texts of the era included cautionary notes, that the spectroscope cannot 
be used alone in forecasting, but that it can be a useful adjunct for weather 
studies.  There are instrumental limitations:  The very narrow slit of a 
spectroscope is likely to become dusty in a pocket instrument, and dust can 
obscure the bands.  Variations in the rainband are introduced by the differing 
widths of the slit, the altitude above sea level during use, the time of day, 
the part of sky viewed, & especially by the temperature.   Prisms with less 
dispersion result in a stronger band; and a longer slit (wider aperture) 
provides a stronger band.
   Observing rain bands is not easy, the changes are subtle and difficult to 
compare over time.  Most users found the subjective estimations of color 
shading to be very difficult, and the use of the spectroscope as a 
meteorological tool faded with the 20th century.
----------

REFERENCES.

Austin, Jill. A Forgotten Meteorological Instrument: The Rainband Spectroscope. 
Bulletin of the Scientific Instrument Society 1 (1983) 9-12.

Browning, John. How to Work with the Spectroscope. London: Browning, n.d.

Capron, J. Rand. A Plea for the Rainband. The Observatory 5 (1882) 42-47,71-77. 
(Reprinted from 'Symons Monthly Meteorological Magazine'.   Republished, with 
remarks, 1886, as 'A Plea for the Rainband, and the Rainband Vindicated'.)

Cory, F.W.  How to Foretell the Weather with the Pocket Spectroscope. London: 
Chatto & Windus, 1884.  (86pp)

Jennings, Donald and Pedro Sada.  Water in Betelgeuse and Antares. Science 279 
(6 Feb. 1998) 844-847.

Konkoly, Nicolaus von. Handbuch fuer Spectroscopiker im Cabinet und am 
Fernrohr.  Practische Winke fuer Anfaenger auf dem Gebiete der Spectralanalyse.  
Halle: Wilhelm Knapp, 1890.

Peterson, Thomas.  The Zealous Marketing of Rain-Band Spectroscopes.  
Rittenhouse 7:3 (May 1993) 91-96.

Schellen, H. Spectrum Analysis. New York: Appleton, 1872; pp178-184.  London: 
Longmans, Green, 1885; pp280-286.  (translated from German)

Smyth, Charles Piazzi.  Spectroscopic Previsions of Rain with a High Barometer. 
Nature 12 (1875) 231-232.

Smyth, Charles Piazzi. Of Spectroscopy, viz., Meteorological Spectroscopy in 
the Small and Rough.  Edinburgh Astronomical Observations 14 (1877) 29-34. 
---------

FOR FURTHER RESEARCH.

Bell, Louis.  Rainband Spectroscopy. American Journal of Science & Arts (3rd 
series?) 30 (1885) 347-354. (formerly Silliman's journal)

Cooke, Josiah.  On the Aqueous Lines of the Solar Spectrum. American Journal of 
Science 11 (1866) 178-184.

Donnelly, J.F.D.  A Meteorological Spectroscope.  Nature 26 (1882) 501. 

Smyth, C.P. The Warm Rainband in the Daylight Spectrum. Nature 14 (1876) 9.

Smyth, C.P. Rainband Spectroscopy. Nature 22 (01 July 1880)194-195.

Smyth, C.P. Spectroscopic Weather Discussions. Nature 26 (1882) 551.

Smyth, C.P.  Astronomical Register, September and/or October 1877
-------------

home page:    http://www.europa.com/~telscope/binotele.htm