A Slitless Spectrograph

Flash Spectrum
Flash spectrum of the solar chromosphere, Nantucket Airport, March 7th, 1970

Extracted from: May, 1970, Sky and Telescope, Page 318 & ff.

A Slitless Spectrograph for the Flash Spectrum

IN 1963, after reading S. A. Mitchell’s exciting book on solar eclipses [1], I became especially interested in his detailed descriptions of the flash spectrum of the sun’s chromosphere, which is observed for a few seconds at the beginning and ending of totality. I constructed a small concave-grating spectrograph and took it to that year’s eclipse at Stetson, Maine. But the equipment was never used, for the very thin crescent of the sun dissolved in clouds five seconds before the second contact.

Photo: A. W. Doolittle

This year [1970] I tried again, taking the spectrograph mounted on my 4-1/4-inch f/28 reflector to Nantucket Island. Spectra were obtained at second contact, at mid-totality, and at third contact, the last being shown in color, four-times enlarged, in the center of this issue. A good flash spectrum is not only colorful but very informative scientifically – an important means for studying the composition and physical properties at different levels in the sun’s atmosphere.

What I needed was a compact instrument that could be attached to a portable, clock-driven equatorial [4] and pointed directly at the sun, eliminating the need for the coelostat arrangement professionals use for their larger equipment (see, for example, pictures on pages 280 and 284 of this issue). At the same time I sought good spectral resolution and large plate scale, in order to record many lines in the flash spectrum instead of just the few brightest ones revealed by direct-vision spectroscopes.

In Amateur Telescope Making—Book Three, Strathmore [2] R. B. Cooke and Robert A. Wilson give a complete geometrical treatise on the Wadsworth mounting for a five-foot laboratory spectrograph. This compact arrangement includes a collimating paraboloidal mirror and slit, without which the light source would have to be an infinitely distant slit. However, at an eclipse this requirement is essentially met by the thin crescent of the solar chromosphere produced by the advancing moon.

Thus, my spectrograph has the great advantage of only one optical element, the concave (spherical) diffraction grating, which focuses monochromatic images of the chromosphere directly on the strip of 35-mm. film that is held at the parabolic focal surface. The instrument’s size, configuration, and performance are completely determined by the grating constants. Replica gratings are perfectly acceptable for most amateur and laboratory purposes, and are offered by Edmund Scientific Co., Central Scientific Co., and others for less than $100.

I chose Edmund’s “instrument-quality concave grating,” which is made of glass 3″ in diameter that has been optically figured and aluminum coated. Its radius of curvature is 1,000 millimeters, and an area 1-1/2 in square is ruled with 15,000 lines per inch. It forms a solar image about five millimeters (1/5 in) in diameter, which easily fits in the width of 35-mm. film and allows room for coronal images. When the grating arrived, I measured its radius of curvature, finding it 10 mm. longer than advertised; my final design was adjusted for the difference in focal length.

The Specifications

Attention was then turned to the geometry of the Wadsworth mounting. The formula
.                                                         nλ = s(sin i + sin θ)           (1)
gives the relation between i, the angle of incidence of the light on the grating, and θ, the angle of diffraction, both measured from the normal (perpendicular) to the grating, as shown in the diagram. Also, n is the order of the spectrum (always a whole number), λ the wavelength in centimeters, and s the spacing between the grating lines, in centimeters.

Because a compact instrument giving maximum image brightness was required, I was limited to the use of the first-order spectrum, so that n was fixed as unity.

The dispersion of the spectrum, expressed in angstroms per millimeter along the focal surface, is given by 107 s/nf, where f is the focal length of the grating (half the radius of curvature), or 505 millimeters for my grating. For s, I used 1/590 millimeter, corresponding closely to 1/15,000 inch, which gave the dispersion as 33.6 angstroms per millimeter.

The spectral range that I wanted to cover lay between 3000 angstroms in the ultraviolet and 8000 in the infrared. This would occupy about 150 millimeters or 6″, that is, 3″ on either side of a central wave length of 5500 angstroms in the green region.

To place this central wavelength of the spectrum upon the grating normal, we set θ = 0 in Equation 1, which becomes
.                                                     sin i = λ/s.            (2)
Expressing- both lengths in millimeters, λ is 5.5 X 10-4 and s is 1/590. We obtain sin i = 0.3245, and  i = 18° 56′.

Using this value of i in Equation 1, we can calculate the values of θ corresponding to 3000 and8000 angstroms. We find that these extremes are symmetrically placed 8° 29′ on either side of 5500 angstroms.

This last calculation is important to be sure that there is enough clearance between the incident rays and the red end of the spectrum to accommodate mechanical and structural elements, such as shutter, film spools and pockets, and the camera walls. For my spectrograph, the clearance is 18° 56′ – 8° 29′ = 10° 27′. Had this been insufficient, I would have set a longer wavelength on the grating normal, obtaining a larger angle of incidence.

The focal curve is not an arc of a circle (a parabola), but is defined by the relation
.                                             D=R cos2 θ/(cos i +cos θ),       (3)
where D is the focal distance at the diffraction angle θ, and R is the radius of curvature of the grating. We can lay out the focal curve by calculating D for selected values of θ, preferably employing logarithms for accuracy.

Mechanical Construction

  Completion of the calculations permitted me to design the spectrograph itself: a lightproof box with suitable covers, access openings, grating support, and film support. In addition, baffles, screens, and light traps were placed to eliminate spurious reflections and stray light. The entire assembly was painted flat black on the inside.

S+T Pg (2)The film was supported at its edges by two strips of Masonite 6″ long and 1/8” thick, filed accurately to fit the calculated parabolic curve. The top and bottom runners were cut and filed together, for a perfect match. Slotted holes in the film support brackets permit slight shifts for the focusing adjustment. A film transport arrangement was devised to fit standard 35-mm. cassettes for winding and rewinding with slotted cranks inserted from the outside. The shutter must be large enough to illuminate the grating area fully, with additional clearance for the field of view (2° in this case). Shutters of large aperture are hard to find, but I came across an old 2-1/2” bulb-operated Packard shutter which worked well after cleaning and oiling. The shortest exposure that I can manage with it is about 1/5 second.

At an angle of reflection equal to the angle of incidence, the source produces a bright zero-order image. To absorb this, I made a black-velvet light trap, as shown in the diagram. But the trap can be opened to allow the strong direct image of the sun to fall on a target, thus providing a convenient means for aiming the spectrograph without opening the film plane access door. The grating holder should be rigidly mounted but adjustable, as in the design by Cooke and Wilson. All but the ruled area should be masked to reduce the intensity of the direct image. To preserve the grating surface, it is essential that the grating cell be fitted with an air- and dust-tight cover.

Provision must be made to set the film curve exactly in focus, not only at the center wavelength on the grating normal but at the other wavelengths along the spectrum. This requires, in the workshop, setting up a collimator and a bright, sharp source such as a slit illuminated by an arc. At the focus of a good 2-inch f/20 achromat I put an adjustable slit like one described by John Strong in ATM-3. To be sure the slit was exactly at the focus of the lens, I pointed the telescope at the star Sirius and moved the slit along the optical axis until the objective darkened uniformly as the star drifted past the optical
axis (as in a knife-edge test).

.                     Disassembled camera                                                    Collimator
Then the collimator was aimed at the grating and the slit illuminated by an arc, one carbon of which had been soaked in a mixed salt solution (NaCI, KNO3, or SrNO3). This produced a rash of bright lines all along the visible spectrum. Using a strip of frosted acetate as a ground glass, and setting the slit for best apparent parallelism with the grating rulings, I used a jeweler’s loupe to focus each section of the film curve. That is, I adjusted the film holder for best focus at each point along the spectrum; the sodium D line was easily resolved into its two components (0.2 mm. separation).

Not apparent when focusing with a slit is an astigmatism which is zero on the normal and increases to either side, becoming very noticeable in the zero-order image. It causes the crescent cusps and other eclipse features not essentially parallel with the rulings to appear somewhat out of focus at the ends of the spectrum. Physicists use cylindrical lenses and other means to overcome this astigmatism, but such a refinement is not necessary for the flash spectrum.

Observing the Eclipse

  I loaded two 20-exposure 35-mm. cassettes to permit trying out two kinds of film at the same eclipse. The leaders were taped together before loading; after loading and closing the camera the splice and a predetermined length of film were wound into one cassette to draw in an unexposed length of film from the other.

After this exposure the winding was reversed and the splice moved from the first cassette into the second for subsequent exposures on the film thus drawn from the first cassette. I did not use 36-exposure cassettes because these might make this operation difficult. I made sure to return the splice to the grating normal before opening the camera for removal of the cassettes.

I was unable to find a complete record of grating constants, exposure time, and film speed for some previous flash spectrum photograph. The speed of the grating is f/Ae, that is, the focal length divided by the equivalent aperture. The latter is Ae = (4As/π)1/2, where A is the actual grating aperture.

However, this calculation is confused by the fact that only part of the reflected light is concentrated into any given order. My grating has an advertised efficiency of 50 to 75 percent and a blaze wavelength of 4000 angstroms, meaning that 50 to 75 percent of the reflected light is concentrated in the first order and further concentrated locally in the 4000-angstrom region.

I decided to try about twice the exposure times recommended for use with a 40° objective prism [3], since substantially all of the light leaving a prism falls within its only spectrum. The exposures used this year on Nantucket Island were 1/5 second on Ektacolor-S (ASA 100) for the flash spectrum at second contact and 1/3 second on High Speed Ektachrome (ASA 160) at third contact (see center pages). Also, a 1-second exposure was made on Ektacolor-S at mid-eclipse, to record the coronal emission.

Before the eclipse the direction of the moon’s relative path was calculated, in order to orient the instrument so that the rulings would be tangent to the midpoint of the chromospheric crescent; this would produce crescents set vertically along the length of the spectrum. However, at Nantucket we were a substantial distance north of the eclipse’s central line, and the crescents could not be made to appear this way at both contacts.

[My assistant and friend Frank Dow manned the hour-angle and right ascension slow motion adjustments to be sure the sun’s image remained centered by cancelling accumulating errors in the motion of the weight-driven clock.]

Without practice from previous eclipses, and in the excitement of the moment, it is very hard to determine exactly when to operate the shutter. The flash lasts only the few seconds it takes the moon’s limb to traverse the solar chromosphere. I placed a small 5,000-line plane transmission grating (Wabash Instruments and Specialties, Wabash, Indiana) over one objective of a pair of binoculars, oriented so the rulings were perpendicular to the moon’s path.

With this, the Fraunhofer absorption lines (crescents) could be seen darkening and sharpening as the “slit” of the photosphere narrowed in the moments before totality. Fleetingly the crescent cusps gleamed and finally the chromosphere’s bright-line spectrum flashed into view as the bright arc of the photosphere was fully extinguished.

  At this moment, the stopwatch was started, to provide a countdown for the recurrence of the flash spectrum at third contact, two minutes later. Of course, with good time-signal reception and an accurate prediction of the duration of totality, second and third contacts could be timed without visual watching, but surely one would not want to trade the beautiful spectacle of the sudden appearance of the flash spectrum for the small cost of a plane transmission grating.

343 South Ave.
Weston, Mass. 02493

[1] Eclipses of the Sun, Samuel Alfred Mitchel, Columbia University Press,  New York, Fifth Edition, 1951, 445 pages.
[2] Amateur Telescope Making – Book Three, Edited by Albert G. Ingalls, Scientific American, Inc., 1953, 644 pages.
[3] Solar Eclipse Photography for the Amateur, Eastman Kodak Co., Rochester, New York, Pamphlet No. AM-10, 1968.
[4] In 2017 the spectrograph and equatorial mounting (tripod with clock drive) were donated to Cornell University’s Fuertes Observatory, in Ithaca, NY.


Total Solar Eclipse of March 7, 1970

In the mid-1960s I began designing and building a spectrograph to photograph the flash spectrum of the total solar eclipse of March 7, 1970. The resulting pictures – several of which were published nationally – appear below, along with background information and work-in-progress shots.

(For the full story of my trip to Nantucket in March 1970, see this post: Astronomy, An Adventure.)


Photos taken with mirror of 3,000mm focal length and 10.8cm aperture by W. Atkinson at Nantucket airport.

Flash Spectrum of the Solar Chromosphere

To describe the appearance (of the flash spectrum) in 1870 to accompany my own photo 100 years later (below, taken in 1970), one cannot do better than to quote from the words of the discoverer, Charles Augustus Young.

An excerpt from S. A. Mitchell’s Eclipses of the Sun, Columbia University Press, 1951, Pg. 104:

“The observation is possible only… at a total eclipse of the sun, at the moment when the advancing moon has just covered the sun’s disc—the solar atmosphere [chromosphere] of course projects somewhat at the point where the last ray of sunlight [from the brilliant photosphere] has disappeared. If the spectroscope (visual in this case) be then adjusted with its slit tangent to the sun’s image at the point of contact, a most beautiful phenomenon is seen. As the moon advances, making narrower and narrower the remaining sickle of the solar disc, the dark (Fraunhofer) lines of the spectrum for the most part remain sensibly unchanged, though becoming somewhat more intense. A few, however, begin to fade out, and some even turn palely bright a minute or two before totality begins. But the moment the sun is hidden, through the whole length of the spectrum, in the red, the green, the violet, the bright lines flash out by hundreds and thousands, almost startlingly, as suddenly as stars from a bursting rockethead, and as evanescent, for the whole thing is over in two or three seconds. The layer seems to be only something under a thousand miles in thickness, and the moon’s motion covers it very quickly.” (C. A. Young).

From the centerfold (pp. 308-309) of Sky & Telescope, May, 1970, Vol. 39, No. 5. © Sky Publishing Corporation 1970

Flash Spectrum, March 7, 1970
The flash spectrum at the conclusion of totality (third contact) on March 7, 1970, by William C. Atkinson and Frank B. Dow, Jr.

Flash spectrum
The arcs are emission lines of the solar chromosphere, conspicuous ones being the D3 line of neutral helium in the yellow (between red and green) at 5876 angstroms and the H and K lines of ionized calcium at 3968 and 3934 in the violet (here rendered blue). The Balmer lines of hydrogen extend from Hà in the red (overexposed) at 6563… In the green at 5303 angstroms, the ringlike image is the corona itself in the light of iron atoms ionized 13 times. Chromospheric lines from within the ring are from the element magnesium.

Sky & Telescope

See also issue cover photo above (left panel) and article in “Gleanings for ATMs”,
A Slitless Spectrograph for the Flash Spectrum (pp. 318-323).

This photograph appears in several astronomy text books (cf. Jay Pasachoff), in NASA’s A New Sun: The Solar Results from Skylab (Chapter 2, p. 31), Geo magazine, and was displayed in the “Hall of the Sun” at the Planetarium of the New York City Museum of Natural History—until its renovation ca. 1999.

(For the full story of my trip to Nantucket in March 1970, see this post: Astronomy, An Adventure.)

Spectrograph and Clock-driven Equatorial by William C. Atkinson

Spectrograph and collimator (construction on my father’s handmade maple workbench)

Carbon arc spectrum on film plane






Astronomy, An Adventure (1925-present)

Irving Porter Church (circa 1920)

My grandfather Irving Porter Church had a small refracting telescope equatorially mounted on a pedestal in his backyard at 9 South Avenue in Ithaca, New York. I can remember it. Somewhere there is a small photo of him in a black frock coat standing beside it. In 1923 the 12-inch refractor of the Fuertes Observatory at Cornell University was named after Prof. Church, then retired chair of the Civil Engineering Department.

In Manhattan, on Saturday, January 24, 1925 (about a week after I was born), there was a total eclipse of the sun. The southern limit of totality was found later that day to have been at 97th Street; observers having been stationed at every other block from 72nd to 135th Streets in order to make the determination. We lived then on West 113th Street, just within the zone where totality lasted only a few seconds.

New York Times Rotogravure Jan 1925

According to my father people had gathered outdoors despite the bitter cold and as the sun reappeared after the brief spectacle a burst of applause rolled across the rooftops of the city. My father later framed its picture cut from the rotogravure section of the New York Times.

I remember, as a child of less than five, a night in Ithaca (probably in August) during which the grownups were to stay up until after midnight to watch a meteor shower (probably the Perseids). Despite a small tantrum I was packed off to bed as too small to stay up that late and I never saw a meteor shower until decades later.

My father, being an engineer and a scientific sort, passed on to us as children a rudimentary interest in astronomy. In the ’30s, when we were ten, knowledge of the Universe was only a fraction of what it is today and didn’t extend in much detail beyond the Milky Way and the more distant galaxies and nebulae: Messier’s objects were still a mystery. The size and age of the Universe was essentially unknown and Edwin Hubble had hardly yet published his theory demonstrating its uniform expansion. We learned that the Milky Way was our own galaxy seen edge-on and came to recognize the Dippers, the Pole star, Orion and the Pleiades in winter, Lyra and Cygnus in summer. Around the house we had some elementary astronomical star charts and texts most of which I had eventually read.

In July of 1932 there was to be a total eclipse of the sun visible on Cape Cod about noon, but not quite total in Boston—which was just west of the central line which swept south from Maine, over the tip of the Cape, and thence out to sea. My father had made reservations for the family on the Provincetown ferry. The ship was crowded with eclipse goers and hawkers offering smoked glass eye protection at outrageous prices. We had our own rectangular panels of heavily exposed photographic film (metallic silver was the agent) sandwiched between glass plates and taped around the edges. They had been made by my grandfather Church.

We climbed to the top of the Pilgrim Tower in time to watch the eclipse from the belvedere. I remember only the brief phase of totality, the dark moon suspended high on the meridian surrounded by the ethereal halo of the sun’s corona.

My interest in astronomy took a holiday until 1944 when I found myself in the Army Air Force in aerial navigation school at Selman Field in Monroe, Louisiana. That spring and summer we flew all over the southwest in twin-engine Beechcraft [AT-7] Navigation Trainers. We were learning pilotage (watching the ground and comparing it to a map), radio navigation (intersecting and following fixed beams from ground transmitters shown on a map), air-plot (navigating “blind” by compass and airspeed through a motionless mass of air—as though there were no wind—and applying an overall averaged wind correction vector at the very end), and celestial navigation (by the sun, moon, Venus, and the stars). [LORAN had been newly deployed but we weren’t exposed to it until later training in Florida and it had not yet been extended to the western Pacific, where I flew combat missions in 1945.]

Much of the mathematics for solving the spherical trigonometry required of celestial navigation could be simplified by using tables prepared by the U.S. Naval Hydrographic Office (HO). The tables were hardbound in books and our particular method was designated HO-218. There were others suitable for various uses (HO-214, etc.) and one tedious, from scratch, by-hand method, requiring no tables, called the Ageton Solution—which we had to memorize. The HO tables were prepared in advance for a fixed set of twenty-two bright stars more or less evenly distributed over the celestial sphere so that, anywhere on earth at any time of night, one could identify and sight on at least three bright stars so spatially distributed as to permit three lines-of-position to be calculated for plotting on the chart. Thus I learned the names and positions of twenty-two stars many of which I would otherwise probably never have known of.

Long after the War my wife Crissy gave me, as a wedding present, an equatorially mounted four and one-quarter inch Newtonian reflector set on a simple tripod. I became fascinated with its possibilities for observing and photography and set about seeing what I could of the Manhattan sky and of the velvet darkness of that same sky in Winchester, Connecticut where Crissy’s family had a country house—“Windrush.”

Soon I was haunting the general astronomy shelf at the New York Public Library and, through the next year or so, read every book more or less in sequence down the length of the public shelf. I learned about making better telescope mounts, making parabolic mirrors, about solar eclipses and famous expeditions to study them, about transits of Mercury and Venus across the face of the sun, about astronomical photography, and the history of the great discoveries of the past. I read all three volumes of “Amateur Telescope Making” and most of Jenkins & White’s “Fundamentals of Optics.” I was hooked.

Month after month (with the exception of the mirror), working in the bedroom and at my friend Lambert Mazzoni’s shop in a loft south of Houston Street, I gradually rebuilt the telescope and its insubstantial tripod ending with a versatile and sturdy mount with setting circles (right-ascension and declination) and driven by simple weight-driven clockwork to cancel the relative motion of the earth’s rotation during long viewing and photographic sessions. Lambert had a lathe and a drill press at the SoHo shop where I could turn wooden, Masonite, and aluminum parts and make simple cameras and accessories. By working at night I could use the kitchen of our apartment on West 108th Street and later on West 116th Street as a darkroom.

Transit of Mercury

61110701_TransitMy first real foray into photography was to capture some successful exposures of the transit of Mercury across the face of the sun on the seventh of November, 1960—photos taken from the roof at 300 W 108th St. I used the direct solar image of the 4-1/4″ reflector at a focal length of 3,000mm, a homemade camera box with a 2-1/4 x 3-1/4 roll film back, and Kodak Autopositive (super contrast) film slips cut to size from a larger sheet. The raw film had to “reversed” by exposure to bright light. After the beginning of the transit (which lasted about three hours) I could make an exposure, run down to the darkened (windowless) kitchen to develop the film, and skip back to the roof for another exposure with altered timing. Mercury was the tiniest of black dots against the image of the sun, much smaller than sunspots nearby (one of which was gigantic). Later the best negatives were printed with a simple enlarging rig made from an old wood framed bellows camera. There was a problem with “limb darkening” whereby the edges of the solar image were significantly less bright than the center. By winding up, and then letting spin down, an empirically adjusted dodge—an internally toothed annulus of cardboard suspended from an axial black cotton thread just above the projected image during the exposure—I could virtually eliminate the unwanted effect. [The 60 degree threads were moving and thus blurred, and the central thread was out of focus.] The best of these results was published in Sky & Telescope magazine in January of 1961 and reprinted in the 1962 McGraw-Hill “Yearbook of Science and Technology” (under “Planet, Mercury”).


I saw my first bright comet (Mrkos 1957d) hanging over the summits in a purple, crystal clear and darkening sky in the Tetons in August of that year. We had to run several hundred yards east, away from the peaks, in order to raise it above the summit ridges. I thought how much fun it would be to be able to photograph (or even to discover) such a one.

Iyeka Seki (1965f)

My attempts at comet photography began in 1959. At first with the 2-1/4 x 3-1/4 camera with eyepiece projection (tiny, tiny images bearing gross enlargement), and later with a new camera (independent of the telescope) assembled from an army surplus aerial lens of much longer (250 mm) focal length with a homemade 4×5 plate holder at the back and permanently focussed with a bright star (Sirius) using a knife-edge installed across a small hole made in the ground glass in the film plane. Eventually I added hand operated slow motion knobs to the right ascension and declination circle adjustments and red-light illuminated cross-wires to the guide telescope thus greatly facilitating the removal of driving clock errors in minutes-long exposures. The intermittency of the escapement gave a slightly muddy, halting motion to the clock. Among the earlier photos were Comets Burnham (1959k) from the roof at 300 West 116th Street, Seki-Lines (1962c) from Dobbs Ferry, NY, and later Ikeya (1964f) and Bennett (1969i) from the yard in Weston, Mass.

The best, though, was Comet Ikeya-Seki (1965f) whose tail faintly swept the entire eastern sky from horizon to zenith before dawn at the end of October. With exposures as long as 30 min (in freezing cold) I was able to get some good pictures with the aerial camera lens. I developed the negatives at night in small enamel trays on the ping-pong table in the darkened basement.

Total Solar Eclipses

Eventually I came across S. A. Mitchell’s “Eclipses of the Sun” (Columbia Univ. Press, 1951) wherefrom I became especially interested in detailed descriptions of the flash spectrum of the solar chromosphere; a phenomenon observable for fleeting seconds at the beginning and ending of a total eclipse of the sun. In particular I was captivated by a description of the discovery of the chromosphere by C. A. Young at the 1870 total eclipse in Spain. He was observing the progress of the eclipse with a hand-held monocular on the objective of which he had mounted a plane diffraction grating:

“As the moon advances, making narrower and narrower the remaining sickle of the solar disc, the dark [Fraunhofer] lines of the [continuous] spectrum [of the waning photosphere] for the most part remain sensibly unchanged, though becoming somewhat more intense. A few, however, begin to fade out, and some even begin to turn palely bright a minute or two before totality begins. But the moment the [photosphere] is hidden, through the whole length of the spectrum—in the red, the green, the violet—the bright lines flash out by the hundreds and thousands almost startlingly; as suddenly as stars from a bursting rocket head, and as evanescent, for the whole thing is over in two or three seconds. The layer [the chromosphere] seems to be something under a thousand miles in thickness, and the moon’s motion covers it very quickly.”

Extraordinary! That was something I had to see, and the more I thought about it, something I had to photograph. It seemed, after some research, that it was something no amateur had yet done. So I set out to design and build a slitless (Rowland mounting) spectrograph; one that I could mount on my clockwork driven equatorial. (For the detailed description of the spectrograph see my article in Sky and Telescope, May 1970, “Gleanings for ATM’s,” p. 318). Usually, to obtain the spectrum of extended objects, a slit is required in the camera itself, but the thin crescent of this object acts as its own slit.

A total eclipse of the sun would be visible in Maine on Saturday, July 20th, 1963.

Slitless Spectrograph

A design began to take shape on my drawing board where I worked at Speed-Park (things were a bit slow), and on weekends at Lambert’s loft in SoHo, and in the bedroom of the apartment on West 116th Street. I combed the junk shops on Canal Street for odd parts—my most serendipitous find: an old bulb-operated Packard shutter of the same 2-1/2″ aperture as the concave diffraction grating of 20″ focal length on whose aluminized surface was a 1-1/2″ square area ruled with 15,000 lines per inch. I had essentially finished the construction by the time we moved house from New York to Boston in June of 1963.

That spring Sky & Telescope published my illustration of the sky at totality showing the stars and planets visible during totality.

The Sky at Totality (C) 1963 William C. Atkinson

Then came final preparations in Cambridge and in my boyhood workshop at 85 Ledgeways in Wellesley Hills. Eventually I spent a last evening on the banks of the Charles River, the spectrograph aimed at the copious neon signage around MIT across the dark river in order, with a jeweler’s loupe, to confirm the focus across the entire length of the parabolic film arc. On the afternoon of the 18th with everything loaded into a new Ford van I headed alone for Maine to allow a day to find a good site and to set up. My friend John Thornton agreed to meet me on eclipse day to help out. I slept on the floor of the van amid the equipment.

Sky and Telescope had predicted the best chance for clear skies at a site on an open hillside southeast of Pleasant Lake in Stetson Maine and virtually on the central line—for the longest possible period of totality. The central line swept down out of Quebec, across central Maine and out to sea at Cadillac Mountain on Mount Desert Island where there was fear of morning fog.

John joined me on Saturday and we spent the morning checking over what I had set up the day before. The morning was nice; clear with drifting fair-weather clouds. Toward noon—the time of the eclipse—the cumuli increased somewhat but everyone (there was large crowd of astronomers and hangers-on strewn across the hillside) seemed optimistic. First contact came—the first nick out of the sun’s disk by the advancing moon—and we then had about an hour. Gradually the clouds billowed larger and filled a greater portion of the sky obscuring the sun for anxious minutes at a time. The sun would reappear (cheers) only to disappear moments later (groans) behind another majestically advancing mass of cloud. The time to second contact (my crucial instant) dwindled to minutes. The world was darkening rapidly; we began anxiously to guess the time when certain blue-sky “holes” would happen along. The moon now covered almost the entire disk of the sun.

A minute to go; sun in the clear. I had a small plane transmission grating taped to one objective of my binoculars and I could see the dark Fraunhofer crescents beginning to condense out of the continuous spectrum of the photosphere. Hand on the shutter bulb, poised, tense. The dark crescents now sharpening, sharpening, scant seconds to go to the flash and then… All dissolved into grayness and faded into virtual night. The cloud took so long to pass that no one even saw the evanescent solar corona, except fleetingly through momentary thinnings in the mist.

The rest of the afternoon was a beautiful summer’s day under a blue sky studded with lamb’s wool clouds. In Orono, at the University of Maine, it poured rain; on Cadillac Mountain, perfectly clear.

The next total eclipse visible on the eastern seaboard would occur on Saturday, March 7th, 1970, the central line passing over Mexico City, leaving the coast at Norfolk, Virginia and grazing Nantucket Island on its way out to sea. This eclipse was part of the 19-year Saros series that included the eclipse of July 1932 that I had seen with my father in Provincetown. For six years the spectrograph gathered dust.

author in 1963
The author in 1963

In 1969 I began a new effort that I came to view as a running battle against Murphy’s Law. I dusted off the spectrograph, telescope, and clockwork and began to prepare for March 1970. I would lie in bed at night conjuring things that could go wrong, each time eventually finding a solution, and in the following days working it out. What if it should rain in the morning; what if there were wind? I arranged the equipment and marked the floor of the 9×9 tent. What if I couldn’t align the polar axis by the North Star the night before or at local noon by the sun? I installed a long surveyor’s compass needle in the main leg of the tripod. What if it were cloudy in the hours before the eclipse? I had calculated the sun’s right ascension and declination at eclipse time so that – once the polar axis of the equatorial mount had been set—I could aim the telescope in advance using only the graduated setting circles. What if the driving clock faltered (as it had in the past)? How could we fine-tune the guiding without a guide telescope (dangerous to the eyes)? And so on and on. Gradually it began to feel as though Murphy could be held at bay.

My friend Frank Dow agreed to accompany me to Nantucket as sorely needed assistance. We set up in the backyard in Weston several times to go through our detailed routine against a stop watch. Frank controlled the guiding using the long tubular lensless sights and the R.A. and declination slow motions to keep the sun’s center always in the cross wires. My plan was to photograph the flash at second contact on Kodachrome film and at third contact on Kodacolor film by taping together the two cassette leaders and winding the film first back into one cartridge then into the other with cranks made from slotted wooden dowels tailored to fit the Kodak 35mm spools. I had a two-spool sample set up that I had used for design, testing, and practice. It was left over from Maine seven years before.

There were to be about ninety seconds of totality between second and third contacts. During this time I would wind film and photograph the corona in color at 3,000mm focal length with the 2-1/4 x 3-1/4 eye-piece projection camera on the 4-1/4 inch Newtonian.

We wondered what chance at all we had for good weather in early March. I had made ferry reservations for the car in December—the reservation clerk wondering why there were so many already booking for the sixth and seventh. The twins (9), Meg and Match, agreed to come too and I arranged a small telescope and camera on a tripod so that Matthew could take some pictures.

During this period the children’s school teachers asked if I would come to their class to explain the eclipse. For the sun I used a blindingly bright photo-flood lamp about four inches in diameter mounted in a large black background sheet and for the corona I painted, on a piece of fly-screen, a diaphanous simulation on the background immediately around the “sun.” A foot or so in front of the sun and slightly larger I mounted a fixed opaque black circle—the “moon.” We set this up on a desk in the front of the classroom at about kid’s eye height and had the children walk slowly across the back of the room. From either side in the back the blinding glare of the “sun” was all that one could see—not the “corona” or even the “moon.” But, as the children walked along, the moon appeared to intrude on the sun’s disk until sudden “totality” in the center when the sun’s corona on the background could then easily be seen. Everybody liked it and I think the kids got the idea. The teacher dragooned the kids into writing nice thank you letters.

Car packing time arrived. We were to leave early on Friday morning and to spend the night at a friend’s house (Julian Everett, an old Dreyfus colleague of mine whom I had visited there before) in Nantucket town. After having packed the instrumental stuff in Weston I assembled a variety of hand tools that I felt might come in handy in general and for emergency—keeping Murphy firmly in mind. Pliers, screw drivers, file, scissors, hammer, hand drill and bits, hacksaw, clamps, wire cutters, wire, tape, glue, oil, and a wood chisel among many others. As I left the workshop I lingered to take a last look around. My eye fell upon a little coping saw and, after a moment of hesitation, I tossed it into the box.

The ferry dock at Wood’s Hole was teeming with activity when we arrived. I had heard that the passage was booked solid and, in spite of earlier confirming phone calls, I was worried a little about the confusion and getting a place in the vehicle line and actually rolling on board. But it sorted itself out and we set sail for Nantucket under sunny skies. Almost everyone on board was on his or her way to see the eclipse.

Nantucket airport was undoubtedly recommended by Sky and Telescope as our observing site. The field, being on the southern side of the island, was just a little closer than the town to the central line of the passage of the moon’s shadow, which passed offshore to the south. There was no practical way of beginning the set up that day, although we did drive down to look over the site. Much pre-eclipse activity was already evident.

Early in the morning we went to the airport. It was hazy but clear and the temperature was above freezing. We found a spot among the many other enthusiasts and the set up went according to plan. We erected the tent, aligned it with north, and set the telescope feet on pre-marked spots. After checking the compass needle in the tripod leg we made small adjustments and precisely at local noon (worked out in advance) we checked the alignment again with the shadow of a plumb bob. In the meantime the telescope and spectrograph were mounted, the image of the sun set on the camera’s ground glass and on the spectrograph’s direct-image cross mark (the “zeroth” order of the spectrum), and the driving clock was started.

Finding a spare minute I set up the small telescope and camera for Matthew; a fixed alignment with a cable release on a 35mm Kodak Retina focussed on infinity.

Fearing degradation from unforeseen light leaks I was loath to load film into the spectrograph unnecessarily early. But now, with about thirty minutes to go, that time had come. I took out the two cartridges, whose leaders had been joined, and set them into the film transport cavities on either end of the 9 inch parabolic film arc. I closed the back and latched it before the insertion the transport cranks from the outside. But, what’s this? They don’t fit? How is it possible that the wooden dowels are now suddenly MUCH TOO LONG?

Seized by panic I was suddenly aware of the baleful presence of Murphy and his inexorable Law: “If something CAN go wrong; it WILL go wrong.” All else was banished from my mind. At home I had failed to try the new film cartridges in the instrument but now I realized that, almost unbelievably in the intervening seven years since Maine, Kodak had altered the 35mm cassette design! What on Earth to do? Only fifteen minutes left! And gradually, as the express bore down, it came to me: the coping saw!

Frantically, in the rapidly gathering darkness I made some simple measurements, marked the dowels and, with the small saw, cut them to the new length and, crucially important, formed the small end-slots required to engage the fins in the cartridge spindles.

Whew! Just in time and with only minutes to go I advanced the film, had Frank check the final alignments, and readied myself to watch the last seconds before second contact with the plane grating on the binoculars.

Poor Matthew had been completely forgotten.

Corona @ 3,000mm fl

It was as it had been in Maine with the difference that the sharpening, darkening Fraunhofer lines suddenly flashed out in brilliant color from red to violet just as Young had described in 1870. I squeezed the bulb and heard the shutter clack open and close in what I hoped would be about a fifth of a second. The flash blazed for about three seconds and faded; totality began; now ninety seconds to its end at third contact.

The corona was beautiful, some stars came out, and everyone in the area fell silent. I wound the spectrograph film, counting crank turns until I knew the film from the second cartridge was in place and that film from the former safely returned to the other cassette. There was time to watch the spectacle and to make and wind several exposures at 3,000mm with the 2-1/4 3-1/4. Then, at third contact, the flash reappeared, facing the other way. I tripped the shutter again.

So great had been the tension that we collapsed to the ground in utter relief. The light returned, intensifying, and the air began to warm. The cassettes we

“eclipse, we made it!”

rewound to expose again the splice. I think we may have opened a couple of beers. Al Doolittle, my friend from J&M, came by and took some pictures. The children wrote in the dust on the side of the van: “Eclipse, we made it.” Murphy had lost.

That afternoon on Nantucket were visible an unusually bright pair of “sun dogs” made by light refracting from ice crystals high in the stratosphere. Perhaps it was an omen; for the film had yet to be developed.

The Kodachrome spectrogram was beautiful (I had had a morbid fear that the processing lab would unthinkingly chop the continuous nine-inch strip into individual slides). I called Sky and Telescope for a meeting and took it in to show Joseph Ashcroft and Dennis Milon. While I sat there S&T decided, for the first time ever, to publish a full color centerfold. The spectrogram stretched across both pages at the top (S&T, May, 1970). They asked me to write an article for the same issue on the instrument’s design and construction.

Flash Spectrum
Flash spectrum of the solar chromosphere

Later I was invited by Dennis Milon to give a lecture at the Harvard Observatory library on the flash spectrum for the Boston ATMs (Amateur Telescope Makers) for which I prepared a fairly elaborate model to demonstrate what Young had seen in 1870.

Ultimately the flash picture was published in several astronomy text books, some other magazines, a NASA publication, and was on display (greatly enlarged) for many years in the “Hall of the Sun” at the Hayden Planetarium in New York City.

Life moved on. I did not realize it at the time but my “career” as an amateur astronomer had ended with the total solar eclipse of March, 1970.


Bill Atkinson, John Reppy
Solar Eclipse in Jackson, Wyoming (August 2017) W. Atkinson and Prof. J. R. Reppy

The telescope gathered dust in the basement in Weston for many years until the time came to break house and move to Cambridge in April 2017. I arranged to donate it to the Fuertes Observatory at Cornell University, and my friend Cornell Prof. John Reppy agreed to shuttle it from Weston to Ithaca.

But it got one more outing, this year, when John and I decided to drive out to the Tetons to view the August 2017 eclipse.