THE STORY OF THE SPECTROSCOPE
Sir Isaac Newton ought really to have been the inventor of the
spectroscope, because he began by analyzing light in the rough with
prisms, was very expert in optics, and was certainly enough of a
philosopher to have laid the foundations of the science.
What Newton did was to admit sunlight into a darkened room through a
small round aperture, then pass the rays through a glass prism and
receive the band of color on a screen. He noticed the succession of
colors correctly--violet, indigo, blue, green, yellow, orange, red; also
that they were not pure colors, but overlapping bands of color.
Apparently neither he nor any other experimenter for more than a century
went any further, when the next essential step was taken by Wollaston
about 1802 in England. He saw that by receiving the light through a
narrow slit instead of a round hole, he got a purer spectrum, spectrum
being the name given to the succession of colors into which the prism
splits up or decomposes the original beam of white sunlight. This
seemingly insignificant change, a narrow slit replacing the round hole,
made Wollaston and not Newton the discoverer of the dark lines crossing
the spectrum at various irregular intervals, and these singularly
neglected lines meant the basis of a new and most important science.
Even Wollaston, however, passed them by, and it was Fraunhofer who in
1814-1815 first made a chart of them. Consequently they are known as
Fraunhofer lines, or dark absorption lines. Sending the beam of light
through a succession of prisms gives greater dispersion and increases
the power of the spectroscope. The greater the dispersion the greater
the number of absorption lines; and it is the number and intensity of
these lines, with their accurate position throughout the range of the
spectrum which becomes the basis of spectrum analysis.
The half century that saw the invention of the steam engine,
photography, the railroad and the telegraph elapsed without any farther
developments than mere mapping of the fundamental lines, A, B, C, D, E,
F, G, H of the solar spectrum. The moon, too, was examined and its
spectrum found the same, as was to be expected from sunlight simply
reflected.
Sir John Herschel and other experimenters came near guessing the
significance of the dark lines, but the problem of unraveling their
mystery was finally solved by Bunsen and Kirchhoff who ascertained that
an incandescent gas emits rays of exactly the same degree of
refrangibility which it absorbs when white light is passed through it.
This great discovery was at once received as the secure basis of
spectrum analysis, and Kirchhoff in 1858 put in compact and
comprehensive form the three following principles underlying the theory
of the science:
(1) Solid and liquid bodies, also gases under high pressure, give when
incandescent a continuous spectrum, that is one with a mere succession
of colors, and neither bright nor dark lines;
(2) Gases under low pressure give a discontinuous spectrum, crossed by
bright lines whose number and position in the spectrum differ according
to the substances vaporized;
(3) When white light passes through a gas, this medium absorbs or
quenches rays of identical wave-length with those composing its own
bright-line spectrum.
Clearly then it makes no difference where the light originates whether
it comes from sun or star. Only it must be bright enough so that we can
analyze it with the spectroscope. But our analysis of sun and star could
not proceed until the chemist had vaporized in the laboratory all the
elements, and charted their spectra with accuracy. When this had been
done, every substance became at once recognizable by the number and
position of its lines, with practical certainty.
How then can we be sure of the chemical and physical composition of sun
and stars? Only by detailed and critical comparison of their spectra
with the laboratory spectra of elements which chemical and physical
research have supplied. As in the sun, so in the stars, each of which is
encircled by a gaseous absorptive layer or atmosphere, the light rays
from the self-luminous inner sphere must pass through this reversing
layer, which absorbs light of exactly the same wave-length as the lines
that make up its own bright line spectrum. Whatever substances are here
found in gaseous condition, the same will be evident by dark lines in
the spectrum of sun or star, and the position of these dark lines will
show, by coincidence with the position of the laboratory bright lines,
all the substances that are vaporized in the atmospheres of the
self-luminous bodies of the sky.
Here then originated the science of the new astronomy: the old astronomy
had concerned itself mainly with positions of the heavenly bodies,
_where_ they are; the new astronomy deals with their chemical
composition and physical constitution, and _what_ they are. Between 1865
and 1875 the fundamental application of the basic principles was well
advanced by the researches of Sir William Huggins in England, of Father
Angelo Secchi in Rome, of Jules Janssen in Paris, and of Dr. Henry
Draper in New York.
In analyzing the spectrum of the sun, many thousands of dark absorption
lines are found, and their coincidences with the bright lines of
terrestrial elements show that iron, for instance, is most prominently
identified, with rather more than 2,000 coincidences of bright and dark
lines. Calcium, too, is indicated by peculiar intensity of its lines, as
well as their great number. Next in order are hydrogen, nickel and
sodium. By prolonged and minute comparison of the solar spectrum with
spectra of terrestrial elements, something like forty elemental
substances are now known to exist in the sun. Rowland's splendid
photographs of the solar spectrum have contributed most effectively.
About half of these elements, though not in order of certainty, are
aluminum, cadmium, calcium, carbon, chromium, cobalt, copper, hydrogen,
iron, magnesium, manganese, nickel, scandium, silicon, silver, sodium,
titanium, vanadium, yttrium, zinc, and zirconium. Oxygen, too, is pretty
surely indicated; but certain elements abundant on earth, as nitrogen
and chlorine, together with gold, mercury, phosphorus, and sulphur, are
not found in the sun.
The two brilliant red stars, Aldebaran in Taurus, and Betelgeuse in
Orion, were the first stars whose chemical constitution was revealed to
the eye of man, and Sir William Huggins of London was the astronomer
who achieved this epoch-making result. Father Secchi of the Vatican
Observatory proceeded at once with the visual examination of the spectra
of hundreds of the brighter stars, and he was the first to provide a
classification of stellar spectra. There were four types.
Secchi's type I is characterized chiefly by the breadth and intensity of
dark hydrogen lines, together with a faintness or entire absence of
metallic lines. These are bluish or white stars and they are very
abundant, nearly half of all the stars. Vega, Altair, and numerous other
bright stars belong to this type, and especially Sirius, which gives to
the type the name "Sirians."
Type II is characterized by a multitude of fine dark metallic lines,
closely resembling the lines of the solar spectrum. These stars are
somewhat yellowish in tinge like the sun, and from this similarity of
spectra they are called "solars." Arcturus and Capella are "solars," and
on the whole the solars are rather less numerous than the Sirians. Stars
nearest to the solar system are mostly of this type, and, according to
Kapteyn of Groningen, the absolute luminous power of first type stars
exceeds that of second type stars seven-fold.
Secchi's type III is characterized by many dark bands, well defined on
the side toward the blue end of the spectrum, but shading off toward the
red--a "colonnaded spectrum", as Miss Clerke aptly terms it. Alpha
Herculis, Antares, and Mira, together with orange and reddish stars and
most of the variable stars, belong in type III.
Type IV is also characterized by dark bands, often called "flutings,"
similar to those of type III, but reversed as to shading, that is, well
defined on the side toward the red, but fading out toward the blue.
Their atmospheres contain carbon; but they are not abundant, besides
being faint and nearly all blood-red in tint.
Following up the brilliant researches of Draper, who in 1872 obtained
the first successful photograph of a star's spectrum, that of Vega,
Pickering of Harvard supplemented Secchi's classification by Type V, a
spectrum characterized by bright lines. They, too, are not abundant and
are all found near the middle of the Galaxy. These are usually known as
Wolf-Rayet stars, from the two Paris astronomers who first investigated
their spectra. Type V stars are a class of objects seemingly apart from
the rest of the stellar universe, and many of the planetary nebulae yield
the same sort of a spectrum.
The late Mrs. Anna Palmer Draper, widow of Dr. Henry Draper, established
the Henry Draper Memorial at Harvard, and investigation of the
photographic spectra of all the brighter stars of the entire heavens has
been prosecuted on a comprehensive scale, those of the northern
hemisphere at Cambridge, and of the southern at Arequipa, Peru. These
researches have led to a broad reclassification of the stars into eight
distinct groups, a work of exceptional magnitude begun by the late Mrs.
Fleming and recently completed by Miss Annie Cannon, who classified the
photographic spectra of more than 230,000 stars on the new system, as
follows:--
The letters O, B, A, F, G, K, M, N represent a continuous gradation in
the supposed order of stellar evolution, and farther subdivision is
indicated by tenths, G5K meaning a type half way between G and K, and
usually written G5 simply. B2 would indicate a type between B and A, but
nearer to B than A, and so on. On this system, the spectrum of a star in
the earliest stages of its evolution is made up of diffuse bright bands
on a faint continuous background. As these bands become fewer and
narrower, very faint absorption lines begin to appear, first the helium
lines, followed by several series of hydrogen lines. On the
disappearance of the bright bands, the spectrum becomes wholly
absorptive bands and lines. Then comes a very great increase in
intensity of the true hydrogen spectrum, with wide and much diffused
lines, and few if any other lines. Then the H and K calcium lines and
other lines peculiar to the sun become more and more intense. Then the
hydrogen lines go through their long decline. The calcium spectrum
becomes intense, and later the spectrum becomes quite like that of the
sun with a great wealth of lines. Following this stage the spectrum
shortens from the ultra violet, the hydrogen lines fade out still
farther, and bands due to metallic compounds make their appearance, the
entire spectrum finally resembling that of sun spots. To designate these
types rather more categorically:--
Type O--bright bands on a faint continuous background, with five
subdivisions, Oa, Ob, Oc, Od, Oe, according to the varying width and
intensity of the bands.
Type B--the Orion type, or helium type, with additional lines of origin
unknown as yet, but without any of the bright bands of type O.
Type A--the Sirian type, the regular Balmer series of hydrogen lines
being very intense, with a few other lines not conspicuously marked.
Type F--the calcium type, hydrogen lines less strongly marked, but with
the narrow calcium lines H and K very intense.
Type G--the solar type, with multitudes of metallic lines.
Type K--in some respects similar to G, but with the hydrogen lines
fading out, and the metallic lines relatively more prominent.
Type M--spectrum with peculiar flutings due to titanium oxide, with
subdivisions Ma and Mb, and the variable stars of long period, with a
few bright hydrogen lines additional, in a separate class Md.
Type N--similar to M, in that both are pronouncedly reddish, but with
characteristic flutings probably indicating carbon compounds.
The Draper classification being based on photographic spectra, and the
original Secchi classification being visual, the relation of the two
systems is approximately as follows:
Secchi Type I includes Draper B & A
II includes Draper F, G & K
III includes Draper M
IV includes Draper N
Pickering's marked success in organization and execution of this great
programme was due to his adoption of the "slitless spectroscope," which
made it possible to photograph stellar spectra in vast numbers on a
single plate. The first observers of stellar spectra placed the
spectroscope beyond the focus of the telescope with which it was used,
thereby limiting the examination to but one star at a time. In the
slitless spectroscope, a large prism is mounted in front of the
objective (of short focus), so that the star's rays pass through it
first, and then are brought to the same focus on the photographic
plate, for all the stars within the field of view, sometimes many
thousand in number. This arrangement provides great advantages in the
comparison and classification of stellar spectra.
When spectroscopic methods were first introduced into astronomy, there
was no expectation that the field of the old or so-called exact
astronomy would be invaded. Physicists were sometimes jocularly greeted
among astronomers as "ribbon men," and no one even dreamed that their
researches were one day to advance to equal recognition with results
derived from micrometer, meridian circle, and heliometer.
The first step in this direction was taken in 1868 by Sir William
Huggins of London, who noticed small displacements in the lines of
spectra of very bright stars. In fact the whole spectrum appeared to be
shifted; in the case of Sirius it was shifted toward the red, while the
whole spectrum of Arcturus was shifted by three times this amount toward
the violet end of the spectrum. The reason was not difficult to assign.
As early as 1842 Doppler had enunciated the principle that when we are
approaching or are approached by a body which is emitting regular
vibrations, then the number of waves we receive in a second is
increased, and their wave-length correspondingly diminished; and just
the reverse of this occurs when the distance of the vibrating body is
increasing. It is the same with light as with sound, and everyone has
noticed how the pitch of a locomotive whistle suddenly rises as it
passes, and falls as suddenly on retreating from us. So Huggins drew the
immediate inference that the distance between the earth and Sirius was
increasing at the rate of nearly twenty miles per second, while
Arcturus was nearing us with a velocity of sixty miles per second.
These pioneer observations of motions in the line of sight, or radial
velocities as they are now called, led directly to the acceptance of the
high value of spectroscopic work as an adjunct of exact astronomy in
stellar research. Nor has it been found wanting in application to a
great variety of exact problems in the solar system which would have
been wholly impossible to solve without it.
Foremost is the sun, of course, because of the overplus of light. Young
early measured the displacement of lines in the spectra of the
prominences, and found velocities sometimes exceeding 250 miles per
second. Many astronomers, Duner among them, investigated the rotation of
the sun by the spectroscopic method. The sun's east limb is coming
toward us, while the west is going from us; and by measuring the sum of
the displacements, the rate of rotation has been calculated, not only at
the sun's equator but at many solar latitudes also, both north and
south. As was to be expected, these results agree well with the sun's
rotation as found by the transits of sun spots in the lower latitudes
where they make their appearance.
Belopolsky has applied the same method to the rotation of the planet
Venus, and Keeler, by measuring the displacement of lines in the
spectrum of Saturn, on opposite sides of the ring, provided a brilliant
observational proof of the physical constitution of the rings; because
he showed that the inner ring traveled round more swiftly than the outer
one, thus demonstrating that the ring could not be solid, but must be
composed of multitudes of small particles traveling around the ball of
Saturn, much as if they were satellites. Indeed, Keeler ascertained the
velocity of their orbital motion and found that in each case it agreed
exactly with that required by the Keplerian law.
Even the filmy corona of the sun was investigated in similar fashion by
Deslandres at the total eclipse of 1893, and he found that it rotates
bodily with the sun. But the complete vindication of the spectroscopic
method as an adjunct of the old astronomy came with its application to
measurement of the distance of the sun. The method is very interesting
and was first suggested by Campbell in 1892. Spectrum-line measurements
have become very accurate with the introduction of dry-plate
photography, and ecliptic stars were spectrographed, toward and from
which the earth is traveling by its orbital motion round the sun. By
accurate measurement of these displacements, the orbital velocity of the
earth is calculated; and as we know the exact length of the year, or a
complete period, the length of the orbit itself in miles becomes known,
and thus, by simple mensuration, the length of the radius of the
orbit--which is the distance of the sun.
If we pass from sun to star, the triumph of the spectroscope has been
everywhere complete and significant. As the spectroscopic survey of the
stars grew toward completeness, it became evident that the swarming
hosts of the stellar universe are in constant motion through space, not
only athwart the line of vision as their proper motions had long
disclosed, but some stars are swiftly moving toward our solar system and
others as swiftly from it.
Fixed stars, strictly speaking--there are no such. All are in relative
motion. Exact astronomy by discussion of the proper motions had
assigned a region of the sky toward which the sun and planets are
moving. Spectrography soon verified this direction not only, but gave a
determination of the velocity of our motion of twelve miles per second
in a direction approximately that of the constellation Lyra. From
corresponding radial velocities, we draw the ready conclusion that
certain groups or clusters of stars are actually connected in space and
moving as related systems, as in the Pleiades and Ursa Major.
Rather more than a quarter century ago, the spectroscope came to the
assistance of the telescope in helping to solve the intricate problem of
stellar distribution. Kapteyn, by combining the proper motions of
certain stars with their classification in the Draper catalogue of
stellar spectra, drew the conclusion that, as stars having very small
proper motions show a condensation toward the Galaxy, the stars
composing this girdle are mostly of the Sirian type, and are at vast
distances from the solar system. The proper motion of a star near to us
will ordinarily be large, and, in the case of solar stars, the larger
their proper motion the greater their number. So it would appear that
the solar stars are aggregated round the sun himself, and this
conclusion is greatly strengthened by the fact that of stars whose
distances and spectral type are both ascertained, seven of the eight
nearest to us are solar stars.
In 1889 the spectroscope achieved an unexpected triumph by enabling the
late Professor Pickering to make the first discovery of a spectroscopic
double, or binary star, a type of object now quite abundant. Unlike the
visual binary systems whose periods are years in length, the
spectroscopic binaries have short periods, reckoned in some cases in
days, or hours even. If the orbit of a very close binary is seen edge
on, the light of the two stars will coalesce twice in every revolution.
Halfway between these points there are two times when the two stars will
be moving, one toward the earth and the other from it. At all times the
light of the star, in so far as the telescope shows it, proceeds from a
single object.
Now photograph the star's spectrum at each of the four critical points
above indicated: in the first pair the lines are sharply defined and
single, because at conjunction the stars are simply moving athwart the
line of sight, while at the intermediate points the lines are double.
Doppler's principle completely accounts for this: the light from the
receding companion is giving lines displaced toward the red, while the
approaching companion yields lines displaced toward the violet. Mizar,
the double star at the bend of the handle in the Great Dipper was the
first star to yield this peculiar type of spectrum, and the period of
its invisible companion is about 52 days. The relative velocity of the
components is 100 miles a second, and applying Newton's law we find its
mass exceeds that of the sun forty-fold. Capella has been found to be a
spectroscopic binary; also the pole star. Spectroscopic binaries have
relatively short periods, one of the shortest known being only 35 hours
in length. It is in the constellation Scorpio. Beta Aurigae is another
whose lines double on alternate nights, giving a period of four days;
and the combined mass of both stars is more than twice that of the sun.
The catalogue of spectroscopic binaries is constantly enlarging; but
thousands doubtless exist that can never be discovered by this method,
as is evident if their orbits are perpendicular to the line of sight or
nearly so. The history of the spectroscopic binaries is one of the most
interesting chapters in astronomy, and affords a marvelous confirmation
of the prediction of Bessel who first wrote of "the astronomy of the
invisible."
Find a star's distance by the spectroscope? Impossible, everyone would
have said, even a very few years ago. Now, however, the thing is done,
and with increasing accuracy.
Adams of Mount Wilson has found, after protracted investigation, that
the relative intensity of certain spectral lines varies according to the
absolute brightness of a star; indeed, so close is the correspondence
that the spectroscopic observations are employed to provide in certain
cases a good determination of the absolute magnitude, and therefore of
the distance. To test this relation, the spectroscopic parallaxes have
been compared with the measured parallaxes in numerous instances, and an
excellent agreement is shown. This new method is adding extensively to
our knowledge of stellar luminosities and distances, and even the vast
distances of globular clusters and spiral nebulae are becoming known.
In fact, but few departments of the old astronomy are left which the new
astronomy has not invaded, and this latest triumph of the spectroscope
in determining accurately the distances of even the remotest stars is
enthusiastically welcomed by advocates of the old and new astronomy
alike.