Wavelengths – Nineteenth-Century Spectroscopy

and the Birth of Modern Astrophysics

Selected Topics From the History of Astronomical Spectroscopy - Theoretical

Development, Instrumentation, and Spectroscopic Discoveries in Astronomy

Andrew Bell [copyrightã2001] - for the Northern California History of Astronomy

Luncheon and Discussion Association (NCHALADA) - January 5, 2002

In 1835, the French philosopher Auguste Comte made a notoriously poor prediction as to the potential scope of future astronomical research:

"[while] we understand the possibility of determining [the stars'] shapes, their distances, their sizes and their movements, we would never know how to study by any means their chemical composition or their mineralogical structure ... in a word, our positive knowledge with respect to the stars is necessarily limited solely to geometric and mechanical phenomena."

The next three-quarters of a century could hardly have proved Comte more wrong.  By the early 20th century, the theory and practice of astronomical spectroscopy had laid the foundation for much of our current understanding of stellar composition and evolution, together with the means for acquiring far more information about the stars' (and other objects) "geometrical and mechanical phenomena" than could ever have been developed on the basis of direct visual observation alone.

1.   Foundation - How Spectroscopy Works

a.   Refraction [prisms]

b.   Diffraction [slits, then ruled gratings]

c.   Interference methods

2.   Beginnings - Optics and Spectroscopy to 1835

a.   Newton's optical experiments [used prisms but failed to observe absorption lines]

b.   Thomas Young's contributions [physiological basis for perception of light, formulated a wave theory]

c.   William Herschel [experimented at infra-red end of the visible spectrum, found heat there]

d.   Joseph Fraunhofer [first significant work; found and accurately catalogued principal absorption lines]

e.   John Herschel [limited laboratory-based experiments during 1820s]

3.   Theoretical Developments - Period of 1835 to 1875

a.   Bunsen and Kirchhoff [publish Chemical Analysis by Spectral Observations in 1860 and articulate basic principles distinguishing between continuous, emission, and absorption spectra.]

b.   The periodic table [first put forth in useful form by Mendeleev in 1869]

c.   The speed of light [reasonably accurate values were already available by early 1800s]

d.   Doppler theory [formulated by Doppler first for light rather than for sound, although his astronomical applications weren't very well thought out].  Fizeau should probably get at least as much credit as Doppler [and he does, or at least he does in France].  Ernst Mach is reported to have arrived independently at many of the same conclusions, too.

e.   Earliest practitioners – widespread applications began almost immediately following publication of Bunsen and Kirchoff's work; e.g., in Italy by Donati (1860) and then Secchi (1863); in England by William Huggins (1863) and then Norman Lockyer (1866); in France by Jules Janssen (1863); and in the U.S. by Lewis Rutherfurd (1863) and then Langley (1867).

f.    Earliest findings [some were correct from the outset (e.g., Lockyer's deduction that helium absorption lines must be those of a previously unknown element; observations of hydrogen emission lines); others were theoretically correct but reported with a bit too much confidence in the precision of the available instruments (e.g., Huggins' early Doppler shift work), and some were just plain wrong (e.g., Henry Draper's "remarkable discovery" of certain "bright lines and bands" in the spectrum of the Sun.]

4.   Precision Instrumentation - The Rowland School

Henry Augustus Rowland (1848-1901) was trained as a civil engineer at Rensselaer in the early 1870s and then taught physics there briefly, prior to accepting his appointment in 1875 as one of the [seven] founding faculty members at Johns Hopkins.  The appointment allowed him to spend a year in Europe prior to the opening of the new university, during which he visited at the laboratories of Maxwell and Helmholtz (among others), and provided $6,000 in funds for the purchase of laboratory apparatus.

a.   Rowland's ruling engine and precision gratings

b.   Rowland's students and the Rowland standards

c.   Rowland's assistant – Lewis Jewell

d.   Michelson and the interferometer

e.   Perot-Fabry corrections published in 1902

f.    St. John (et. al) revised tables of 1928

5.   Photography and Photographic Techniques

Absent the development of photography, astronomical spectroscopy could never have advanced beyond study of the spectrum of the sun and those of the very brightest stars.  Early photographic applications required extraordinarily long exposure times, and this continues to be true for many applications even today.  It's hard to get enough photons in the first place, and then the spectroscope spreads them all out!

a.   Photography and photographic applications to astronomy

b.   Spectrographic applications of photography

6.   Early Astrophysical Results

I am describing the first approximate three-quarters of a century of astronomical spectroscopy as having laid the foundation for modern astrophysics.  By this, I mean to include much of our current understanding of (following Comte), both "the stars' shapes, their distances, their sizes and their movements" and "their chemical composition [and] mineralogical structure," as well as (now leaving Comte behind), the composition, evolution, and physical mechanics of the larger scale universe.  As time permits, we can use this portion of our discussion to talk about how each of several fundamental astrophysical applications were developed and practiced over that time subsequent to the arrival of the necessary instrumentation tools [of the necessary degree of precision] and conceptual frameworks which were prerequisites for productive analysis in each application area, forward to perhaps about 1925:

a.   Physics and Chemistry of the Sun

b.   Spectroscopic Binaries as Dynamic Systems

c.   Local-Scale Physical Mechanics

d.   Spectral Analysis of "the Nebulae"

e.   Stellar Classification and Evolution

f.    Large-Scale Physical Mechanics

7.   Important Observatories and Observing Programs

      The modern scientific community as we might define it today did not emerge in continental Europe until the early decades of the 19th century [using criteria of established scientific journals and availability of formal post-graduate training], and not in the United States until at least 1875.  It was a small place!  Smaller still would be the worldwide community of practicing astronomers and (nascent) astrophysicists, and this would continue to be true for some decades into the 20th century. 

a.   Potsdam

b.   Allegheny

c.   Harvard

d.   Mt. Hamilton

e.   Yerkes

f.    Mt. Wilson

Suggested Discussion Topics

1.   Direct visual observation – (following Comte), do spectroscopic methods count, were we to insist on following the "true science requires direct observation or experiment" principle?

2.   Late 17th century spectral analysis – how much work did Newton really do in this area? Why wasn't he able to find the principal absorption lines, possibly as much as 150 years ahead of the work that Fraunhofer would publish in 1817?

3.   The Stefan-Boltzman equation – why did it take so long for surface temperature to become generally used as a good method for classifying stars?  Blacksmiths would have already understood the relationship between temperature and primary observed colors, so why does it seem like astronomers were so slow to use it?

4.   Henry Rowland and his students

a.   How should Rowland's accomplishments be measured?  There is still a great deal of tribute given to Rowland in the scientific instrumentation literature, but precious little in what he might call the modern literature of "true" science, if he were he alive to look at that literature today.  Is it possible he should be looked at just as an idiot savant, with nothing to his credit beyond the development of the precision ruling engine?  How significant was that invention, anyway?  Michelson's own daughter, in her "The Master of Light" (1973), writes that Michelson devoted years of his life to attempts to improve on what Rowland had done.  And failed.

Rowland has two entries in "Selected Papers of Great American Physicists" (American Physical Society, 1976).  It might be interesting to look at those two papers while discussing this question.  The first one is just an entry that he submitted for the 9th Edition of the Encyclopedia Britannica (____), called very simply "The Screw."  And the second one is his, "The Highest Aim of the Physicist," the address he delivered in 1899 as outgoing first president of the APS.

b.   Lawyers, guns and money – as interests dictate and time allows, it might be interesting to spend a few minutes talking about Rowland's rather conflicted attitudes towards "practical science" in general, and engineering applications in particular.  Which fields on the one hand he feigned to greatly disdain [following Swift, his "two blades of grass" (as cited in Greenberg (1967))], and yet on the other he seems to have trained at least as many students in electricity and magnetism (and the applications thereof) than he ever did in "true" physics, at least if that's what we should call spectroscopy.

c.   Rowland's students – It might be interesting to try comparing the subsequent professional careers of Rowland's doctoral students with those from comparable U.S. graduate programs in physics during the same period.  No major prize winners as far as I know, but a surprisingly large number of them did go on to fairly distinguished careers, in what might seem today to be an unusual variety of different fields.  But I would want to be a little careful about giving Rowland too much credit for these students' later success.  The problem is there just aren't any comparable U.S. graduate programs to compare his against for the period during which he was active.  In fact, there was pretty much no other program, period.  If you were a good student in the physical sciences in this country during the last quarter of the 19th century, your choices were fairly limited.  You could go to Baltimore and work under Henry Rowland or you could buy a boat ticket (e.g., Michelson in the late 1870s), and that was about it. [c.f. "Cottrell: Samaritan of Science" (2nd ed., 1993) for a physical chemistry student's choices in pursuing graduate study at home or abroad, two decades after Michelson].

5.   Precision, accuracy, and 19th century scientific practice –  I think it's possible that in the earlier part of the 19th century, scientists were a little more careful about recognizing the limits of precision inherent in their instruments than their successors a generation or two later.  Later in the century, the practitioners' confidence in their results may have gotten a few steps out in front of the gains afforded by the great leaps forward (remarkable as they were) in the quality and accuracy of the available instrumentation [c.f. Wise (ed.), "The Values of Precision" Princeton (1995)].

6.   Early applications in modern astrophysics – in the next-to-last section of the first part of this outline, I described six fundamental early application areas in astrophysics; briefly:  physics and chemistry of the sun, spectroscopic binaries as dynamic systems, local-scale physical mechanics, analysis of the nebulae, stellar classification and evolution, and large-scale physical mechanics.  This classification system is my own, and I would like to know whether it strikes others as useful.

I have also tried to arrange these six categories more or less chronologically, in order of the approximate historical periods during which it first became possible to study each one productively.  So I would also like to hear any thoughts as to whether the order seems about right.

a.   physics and chemistry of the sun [c.f., Lockyer (England); Langley (Allegheny); see also Henry Draper's "remarkable discovery" of c. 1877; and historical development of this entire subject area over time as presented in contemporaneous basic astronomy texts]

b.   spectroscopic binaries – earliest confirmed finding was by Vogel at Potsdam (Algol, 1889); second was by Antonia Maury at Harvard (b Aur, also 1889).  For early publications in PASP and ApJ, see Keeler (1890) reporting Vogel's findings for Spica (includes good description of expected tidal forces) and Vogel (1903) for e Aur [see either Burnham, Jr. (1978) or Moore (1987) for both Algol and e Aur]

c.   local-scale physical mechanics – see Vogel (1898, 1900) and Hartmann (1900) for early work done in this area at Potsdam; Seares (1918) for discussion of star distributions incorporating both radial and proper motion studies; and Rubin (1995, at p. 422) for those conclusions that William Herschel had been able to draw even before the advent of spectroscopy.  I enjoy showing people the double cluster in Perseus as the single readily-observable "display object" located in an adjacent spiral arm rather than our own [see Burnham, Jr. (1978), Vol. III at pp. 1438-1447]

d.   analysis of the nebulae – first nebula to be studied using spectroscope was NGC 6543 (by Huggins, 1864).  It is a small bright one in Draco of unusually high surface brightness (e.g., just two-fifths of an arc-minute across, but with a surface brightness of nearly mag. 6), so it was a pretty good choice as the first target.  The two bright doubly ionized oxygen lines on either side of 5000 Angstroms popped right out [initially resolved only as a single line, and not explained chemically until 1928 (c.f., "Nebulium")].

Bright emission lines were observed for large numbers of what we now know as the intra-galactic nebula during the late 19th and early 20th centuries.  Those of "the nebulae" that we now know to be extra-galactic were much harder.  There were at least a few people who had it right somewhat in advance of Hubble, though [e.g., Scheiner (1899), Seares (1920), and Oepik (1922), for three fairly good early analyses of the "Great Nebula" in Andromeda].

e.   stellar classification and evolution – It took quite a bit longer to work out enough of the details about absorption-line spectra and begin making productive use of them than it did for bright emission-line spectra and so to determine the chemical composition of at least the intra-galactic nebulae.  There are some pretty good reasons why it took so much longer (just for starters, emission lines are few and bright, while absorption lines come in bunches, and they are dark).

A number of the earliest stellar investigations were of nova, with the first recorded spectroscopic study being Huggins' of T CorB in 1866.  See Burnham, Jr. (1978), Vol. II at pp. 708-714 for a reasonably modern discussion of its first appearance and how it was studied by Huggins and others.  (A somewhat maudlin fictional treatment of the same subject would appear three years later in Appleton's Journal.  See J.B. Bouton (1869) for "Linked to a Star.")  T CorB also proved to be the first example of a recurring nova, with its second appearance early in 1946 resulting in a number of new studies published shortly thereafter.

Virginia Trimble has described Cecilia Payne's single greatest contribution to astronomy (there were others, too) as, "the demonstration that all normal stars have essentially the same chemical composition, dominated by hydrogen and helium."  That certainly would have been one of the very big keys to getting ready for modern studies of stellar composition and evolution.  [See Payne (1925) for the original work, as well as treatment in Haramundanis (ed.) (1996, 2nd ed. at pp. 20-23 (Kidwell) and 159-166 (Payne)].   

f.    large-scale physical mechanics [c.f., Rubin (1995); Trimble (1995b); and Sandage (1999); also, include citations to Adriann van Maanen's now-notorious wrong turn (photographic rather than spectroscopic observation of galactic rotation) together with Harlow Shapley's oft-cited and sorrowful, "I believed in van Maanen's results ... After all, he was my friend" (e.g., as discussed in Haramundanis (ed.); see Virginia Trimble's Note 3 at pp. xx-xxi, as well as pp. 208-210 for Cecilia Payne's version, putting Shapley quote into its historical perspective].

7.   Early astrophysical observatories and observing programs – here, I have just tried to name a "short list" of what strike as the most important of the earliest places where there were well-defined observing programs using spectroscopy.  I came up with Potsdam, Allegheny, Harvard, Mt. Hamilton, Yerkes, and then Mt. Wilson.  And once again, I tried to arrange this list more or less chronologically; not necessarily in order of when they were founded, but in order of the periods during which it looks to me like the most significant astrophysical observing programs were being carried out.

I wanted to be sure that I included at least one observatory that was located outside of the U.S.  I chose Potsdam because so many of the early research papers that I have already looked at came from there.  I'm sure that those fond of France or England could suggest others, and encourage them to do so!

8.   On reading old books and old scientific obituaries – there is neither room nor time for more than a few words on this subject here.  Briefly, I greatly enjoy the former (and find them tremendously useful), and find the latter helpful if read with caution [c.f., hagiography][!]