computers in astronomy

Dating van Gogh's Moonrise

Astronomical sleuths head to France to answer some nagging questions about a 19th-century painting.

By Donald Wo Olson, Russell Lo Doescher, and Marilynn So Olson

shows wheat stacks in a field enclosed bya stone wall and, in the twilight sky, aprominent orange disk partly hidden be-hind a mountain range. Once again, as-tronomical calculations helped eliminatethe ambiguity.

RT HisTORIANS HAVE LONGagreed that Vincent van Goghpainted his spectacular Starry

~ lVight between June 16 and18, 1889, in the town of Saint-Remy-de-Proverice in southern France. Evidencefor pinning down the when and where ofvan Gogh's works comes largely from hisown correspondence with his relativesand colleagues. Sometimes there areother clues within the works of art them-

selves. As astronomers, we have a keeninterest when the paintings' subjects in-clude objects in the sky.

Previously, we identified Venus as theheavenly object in van Gogh's White Houseat Night, painted in northern France(S&T: April 2001, page 34). During ourresearch for that project, we became in-terested in another astronomical paint-ing from Saint-Remy that had an uncer-tain date. This painting, known as F735,

Van Gogh in Saint-RemyAfter suffering a series of medical criseswhile in the town of Aries, van Gogh

moved to the hospital housed in the

Past art historians have been able to determine some details about this painting by Vincent van Gogh, but astronomers were able to pinpoint the

timing of the scene. Once thought to show the setting Sun at Saint-Remy-de-Provence, France, it is now known to be a rendition of a rising Moon.

54

Left: In May 1889, van Gogh moved from

Aries to Saint-Remy-de-Provence to restore his

health. He stayed for a year and created near-

ly 150 paintings, inspired by his surroundings.

Below: The overhanging cliff painted by van

Gogh actually exists to the southeast of the

Saint-Paul monastery in Saint-Remy. The au-

thors determined the altitude and azimuth of

the outcropping by observing the Sun, Moon,

and stars for six days. In this photograph the

Cn Moon is only faintly visible because the Sun"

was well above the horizon.

Saint-Paul monastery at Saint-Remy onMay 8, 1889. By the time of his dischargeon May 16, 1890, he had produced near-ly 150 paintings and 140 drawings -works that reflected his interest in thenatural light of southern France and thecomfort and reassurance he found in na-ture. More than a dozen of these showsimilar views of the enclosed wheat field,with houses visible beyond the wall andthe Alpilles Mountains rising on the

right.lVan Gogh often looked at the sky

from his room on the upper floor of theeast side of the monastery, as he men-tioned in a letter (no.592) of late May1889 to his brother Theo in Paris:

Through the windows with iron bars Isee an enclosed wheat field. ..abovewhich I see the sun rise in all its gloryin the morning.

Another letter (no.593), from the firstweek of June 1889, describes an observa-tion ofVenus:

This morning I looked at the country-side from my window for a long timebefore sunrise with nothing but themorning star which appeared very large.

The view across the enclosed field isthus generally toward the east, and paint-ing F735 therefore depicts the rising (notsetting) of a celestial body.

Sun? Moon? What Month?Much as modern astronomers use Messier

numbers, art historians refer to van Gogh'sworks by the catalog numbers assigned

by Jacob-Baart de la Faille in his 1928compilation, which inevitably containedsome errors, as do all such monumentaland pioneering works. For example, he

listed F735 with the title Sunset and re-marked that the "yellow-orange solar

globe sinks behind the dark blue moun-tains that limit the horizon." A 1937 cat-

alog by Willem Scherjon and W. J. deGruyter altered the title of F735 to RisingMoon (Haycocks) and placed it with theworks from August or September 1889.In his revision published in 1939, de la

Faille titled F735 Moonrise and favored

September 1889. But the most recent(1970) editio? of the de la Faille catalog,

1 Readers can find images of van Gogh's paint-

ings and the full text of his letters at www

.vangoghgallery.com. The Saint- Remy enclosedwheat field also appears in paintings F618, F650,F718, and F737, drawing F1552, and malty others.

completed by a committee of e~ertsafter the author's death, assigns to F735the title Rising Moon: Haycocks and givesthe date as July 6, 1889.

All these scholars have used van Gogh'smany letters to help determine the con-tent and sequence of his paintings, andin this case the correspondence makes itcertain that F735 shows a rising Moon.During the summer of 1889 van Goghmailed an envelope containing two let-ters, one to Theo (no.603) and another(now lost) written to the artist Paul Gau-guin, with a sketch of a reaper cuttingthe wheat in the enclosed field. Vincentdescribed to Theo several completedpaintings and then went on to add:

I have one in progress of a moonriseover the sam~field as the sketch in theletter to Gauguin, but with wheat stacksreplacing the wheat. It is dull yellowochre and violet. Anyway, you will see itsometime soon.

During this same summer van Gogh'shealth took a downturn, and he pro-duced almost no paintings for about sixweeks. Did he paint the moonrise andwrite letter 603 before or after this hia-

tus? Unfortunately, the letter itself doesnot bear either a handwritten date or a

postmark.Letter 603, numbered by Theo's wife

after her husband's death, is consideredespecially troublesome to date. All mod-ern scholars agree that it must be movedin the sequence (that it does not fall be-tween letter 602 of late August and letter604 of early September), but they differon exactly what date to give to letter 603and thereby to the moonrise painting "in

"progress.

Jan Hulsker, an authority on vanGogh's correspondence, noted in 1972,"The change in the sequence of theseletters is not only important from the bi-ographical angle, but a1so for the chro-nology of van Gogh's art for it meansthat the canvases referred to in 603 shouldbe dated two months earlier:' Hulskermoves letter 603 to July 6th and datesthe medical crisis from "about July 8 tillabout the middle of August:' A recentcatalog by Hulsker (The New CompleteVan Gogh, 1996) likewise dates F735 asJuly 6, 1889.

But Ronald Pickvance, curator of theMetropolitan Museum of Art's 1986 ex-

Sky & Telescope I July 2003 55

>.

Eoco~+IIIIra

C

III~GJ+I~0.

E0v

hibition "Van Gogh in Saint-Remy andAuvers:'places the moonrise painting inthe range of July 8-13. Pickvance favorsa date Qf about July 14th for the letter(rio. 603) that describes the moonrisecanvas and July 16th for the onset of thecrisis.

Can we use astronomical methods todate the moonrise painting directly, with-out relying on letter 603?

From left to right: Russell Doescher, Claude Suc, and Donald Olson plan observations in a

Saint-Remy olive orchard. Photograph by Marilynn Olson.

ings from Saint-Remy.We wondered: Does Moonrise accu-

rately depict the landscape near Saint-Paul monastery? Could we find the dou-ble house and the overhanging cliff?What are the azimuth and altitude of thecliff as seen from the wheat field? Did anearly full Moon rise in that direction ona date during 1889?

Possible Dates and QuestionsWe can be certain that van Gogh createdF735 between May 8th, when he first ar-rived at Saint-Remy, and late September,when he mailed the canvas to Theo in abatch of 10 paintings that also includedthe famous Starry Night.

The bright orange orb in F735 mustbe either a full Moon rising near sunsetor a nearly full Moon rising in eveningtwilight, shortly after sunset. Computercalculations for 1889 quickly showed theonly possible dates: May 15-17, June13-15, July 12-14, August 11-13, andSeptember 9-11.

A striking feature of Moonrise is thedistinctive overhanging cliff that partial-ly blocks the disk of the Moon. VanGogh also included, in the foothills be-low the cliff, an unusual "double house"with the appearance of a smaller struc-ture attached to a larger building. Wewere encouraged that these might bereal features of the landscape becausethe overhanging cliff and the double houseappear in so many of van Gogh's paint-

based in Aries. Three of the group'smembers, Claude Suc, Vincent Suc, andBruno Massal, helped us by scoutingpossible observing locations and then ac-companying us during our visit.

Within minutes of our arrival in Saint -

Remy, we were gratified to see that theoverhanging cliff actually does exist tothe southeast of Saint-Paul! Measuringits azimuth and altitude was not simple,in part because Saint-Paul is still a work-ing hospital, with both van Gogh's for-mer room and the enclosed field (now agarden) strictly off-limits. (Tourists areallowed to visit a different room called

Trip to ProvenceTo answer these questions, we traveled to

Saint-Remy in June 2002. Before leavingTexas, we contacted Les Astronomes Am-ateurs Du Delta, an astronomy club

Above: The "double house" of the Moonrise

painting still stands in Saint-Remy. Photo-

graph by Donald Olson.

Left: Van Gogh included the "double house"

and overhanging cliff in several paintings.

This one, designated F617, was painted be-

fore Moonrise, as evidenced by the wheat in

the process of being harvested. Courtesy

Kroller-MiillerMuseum, The Netherlands.

56 July 2003 I Sky & Telescope

..

"Reconstitution Chambre Van Gogh:')More important, a forest of tall pinetrees has grown, obscuring some featuresof the landscape. For example, by explor-ing dirt roads into the forest, we foundthat the double house still stands in asmall clearing about 640 meters (2,100feet) southeast of the monastery, but thisdistinctive building can no longer beseen from locations near Saint-Paul.

The ArIes group had forewarned us ofthese problems and helped us find an ex-cellent observing location -a largeopen field immediately northwest ofSaint-Paul. From here we had a clearview directly over the monastery to theAlpilles beyond. For six days and nightswe observed the Sun, Moon, and stars asthey rose and thereby measured the alti-tudes and azimuths of the peaks andcliffs of the mountain range.

'~

IT

c~""M""

~

"~'

I

Perfect for visual, CCD, and photographic work

Knife-edge baffling produces rich contrastAir spaced, triplet objective yields amazing color correction

Long back-focus accommodates binoviewers, color-filter wheels, all

film and CCD cameras

* Texas Nautical Repair, Inc.

3110 S. Shepherd, Houston, Texas 77098Phone: 713-529-3551 .Fax: 713-529-3108

www. taka hashia merica.com

Narrowing the ChoicesVan Gogh's viewpoint in the enclosedfield was near the north wall, barely visi-ble at the extreme left edge of the paint-ing. We employed topographic maps andaerial photographs to determine that theoverhanging cliff stood 2,680 meters(8,800 feet) from van Gogh's position,and that our observing location was an-other 300 meters farther back and 6 me-ters lower in elevation. Using trigonome-try to make small corrections thatadjusted our observations of altitude andazimuth, we found that van Goghwouldhave seen the outcropping near azimuth126°, that is, about 36° south of due east,and extending between 4'/20 and 43/40 inaltitude. At the latitude of the monastery(43° 47' north), celestial objects risingfrom behind this cliff must have declina .:tion near -21'/2°0 Our computer calcula-tions then showed that from the wheatfield in evening twilight, van Gogh couldhave seen a nearly full Moon rising withthis declination on only two dates in1889: May 16th and July 13th.

Weather records, preserved at the MeteoFrance archives, show that favorable condi-tions prevailed ollboth of these evenings.Very heavy rain fell on May 14th and15th, but skies cleared on May 16th. Theentire' first two weeks of July were rainfree, and the fraction of the sky coveredby clouc;ls decreased during the course ofJUly 13th from 50 to 30 percent.

The colors in the foreground of Moon-rise allow us to eliminate one of thesedates. Shortly after arriving in Saint-Remy in May, van Gogh noted that the

* Adirondack Video Astronomy. Glens Falls, NY * Hands On Optics. Damascus, MD

518-812-0025 866-726-7371* Anacortes Telescope. Anacortes, WA Hardin Optical. 8andon, OR

800-850-2001 877-447-4847

Astronomics .Norman, OK * Oceanside Photo & Telescope. Oceanside, CA800-422-7876 760-722-3348

* 8.T. Technologies. North Miami 8each, FL Optical Guidance Systems. Huntingdon Valley, PA

305-652-3115 215-947-5571* 8oston Telescope. Newton, MA Perceptor .Schomberg, ON

800-608-8439 905-939-2313

* These authorized Takahashi dealers should have a TOA-130 displa\,ed in their showrooms.

Sky & Telescope I July 2003 57

monastery was surrounded by "very, verygreen wheat fields" (letter 591). By themiddle of June, he was painting "a fieldof wheat turning yellow. ..the ears ofgrain baked by the Sun. ..with the warmtones of a bread crust" (letter W12). Inlate June, the artist painted a reaper cut -

ting wheat that was "all yellow" (letterno.597). The Moonrise painting, withreaped wheat in golden stacks, thereforecannot correspond to a date in mid-Maybut fits perfectly with our astronomicallyderived date of July 13th.

Near the summer solstice, nearly full Moons

must have a southern declination and rise in

evening twilight over the Alpilles Mountains

to the southeast of the Saint-Paul monastery.

On only two dates in 1889 did such a Moon

rise exactly behind the overhanging cliff, sev-

eral degrees to the left of the view in this

image from June 23, 2002. Photograph by

Donald Olson.

The Moon in July 2003Throughout 2003, as special events com-memorate the 150th anniversary of vanGogh's birth, there will also be an astro-nomical anniversary. Exactly six 19-yearMetonic lunar cycles have elapsed sincevan Gogh's summer in Saint-Remy.Therefore, the calendar dates of lunarphases in 2003 nearly repeat those of1889. On July 13, 2003, modern ob-servers during evening twilight will see anearly full Moon rise in the southeast,much as it did on July 13, 1889, whenvan Gogh stood among the wheat stacksin the monastery field and captured thescene in his remarkable Moonrise.

Final ChoiceBecause the Moon's disk spent less thantwo minutes passing behind the overhang-ing cliff, we can determine a precise timefor van Gogh's moonrise: exactly 9:08p.m. local mean time2 on July 13, 1889.

Because van Gogh produced approxi-mately one painting or drawing per day,our calculations therefore suggest thatletter 603, describing the Moonrise as inprogress, would date from about July 14,1889 -in agreement with Pickvance'shistorical analysis!

Our topographic observations, com-bined with computer calculations, pro-vide strong evidence that van Gogh wasworking from nature -and not fromsome composite from his memory -

when he created Moonrise; we have anaccurate depiction of both the overhang-ing cliff in the Alpilles and the positionof the rising Moon. The shadows of thewheat stacks are not aligned with theMoon, suggesting to us that van Goghremained in the field as the evening twi-light faded and the rising Moon began toswing toward the southern sky, causingthe shadows to rotate in azimuth.

Furthermore, in an 1888 letter (no.B19) to his artist friend Emile Bernard,van Gogh described his reliance on thenatural world:

1 never work from memory. ...I cannotwork without a model. ..in the matterof form I am too afraid of departingfrom the possible and the true. ..myattention is so fixed on what is possibleand really exists. ..I exaggerate, some-times I make changes in a motif; butfor all that, I do not invent the wholepicture; on the contrary, I find it allready in nature. ...

DON OLSON and RUSSELL DOESCHER teach

in the physics department, and MARILYNN

OLSON in the English department, at South-

west Texas State University. The authors have

previously derived astronomical dates for two

other paintings by van Gogh (October 1988

and April 2001 issues) and a moonrise photo-

graph by Ansel Adams (December 1994 issue).

The authors are grateful for research as-

sistance from Charles Whitney at Harvard

University, Norbert Aouizerats at the Meteo

France archives in Provence, and Margaret

Vaverek at Southwest Texas State University's

Alkek Library.

2 France had not yet adopted modern time zones;

in 1889 Saint-Remy kept local mean time, 19

minutes later than Universal Time.

58 July 2003 I Sky & Telescope