Nephosthe weather behind the paintings
Martin Johnson Heade, Newburyport Meadows: haystacks and harvesting figures on a sunlit marsh, beneath dark clouds, rain shafts at left and diagonal sunbeams.
Martin Johnson Heade, Newburyport Meadows, about 1876–81. Oil on canvas. The Metropolitan Museum of Art, 1985.117. Public domain, Met Open Access.

A study against the weather record

Martin Johnson Heade

Newburyport Meadows

About 1876–81 · The Metropolitan Museum of Art, 1985.117

Rain at the left, sunlight on the hay meadow. Did the harvest season bring both on the same day?

5+rain dates with a clear daytime sky entry in the local hay seasons of 1879 and 1880. The weather type recurs; the painted day stays open.

Rain and light recurPainted day unknown

The recordsThe painting at The Met

The Nephos method

How this study was read

The method in full

  1. What is written

    Nephos reads stacked salt hay, rain shafts at left, sunbeams and a lit marsh as a passing summer shower during salt-haying. This is Nephos’s interpretation of visible features, not a museum statement or a proposed painted day.

  2. Records of those days

    • Newburyport rain and original cloud observationsCastelhun’s daily rain register is complete for July–September 1879 and 1880. The original Form H sheets add daytime upper/lower cloud entries omitted from the CSV.
    • The cloud sheet and its keyThe printed form defines zero as entirely clear. Nephos accepts only explicit zeros in both layers at 7 a.m. or 2 p.m.; night, blank and faint entries do not qualify.
    • Local haying season and its interruptionMass Humanities places cutting and stacking from late June to September; the Herald documents wet weather and high tides stopping Oldtown salt-hay work in late July 1881. The test uses the full months July–September.
    • Newburyport 1991–2020 normalsThe same three months total 10.90 inches and 28.7 rain days ≥0.01 inch at USC00195285. This modern station remains separate from Castelhun’s.
    • Heade’s working methodFulton’s technical study documents development and adjustment of compositions on canvas across other works. It is not a technical examination of this Met painting.

    Not found or not yet searched

    • No dated sketch, letter or outdoor session has been tied to this canvas in the sources checked.
    • The painted viewpoint is unlocated, the historical gauge position estimated, and the daily rain observation hour unknown.
    • The Met’s change from the earlier about 1872–78 proposal to about 1876–81 is unexplained here.

  3. Each claim tested

    • Did rain and clear daylight share hay-season dates? Yes: at least 5 dated records join positive rain with explicit clear daytime sky, among 48 rain dates in 184 days. This conservative minimum excludes partly cloudy and uncertain entries.
    • Does a contemporary witness describe passing showers? A beautiful sunset after the fine showers of Sunday. The 26 August 1878 Herald report is outside the two complete rain seasons and is not added to their counts.
    • Did weather matter to the harvest? The 2 August 1881 Herald explicitly says wet weather and high tides had prevented Oldtown farmers making salt hay for two weeks.
    • What can then-versus-now establish? Historical rain-day counts are 25 and 23; the modern seasonal normal is 28.7. Different sites and two historical seasons do not establish a climate trend.
    • What did Heade arrange? The Met describes reduced foreground detail, a broader and lower horizon and less regular haystacks. This supports a compositional reading, not a claim that he altered observed weather.

  4. Verdict

    At least 5 of 48 rain dates in the local hay seasons of 1879 and 1880 also have a clear daytime sky entry. The Herald records sunset after showers and salt-hay work stopped by wet weather and high tides. These support the weather type in Nephos’s reading; they do not date the painting.

    • Partially agreesThe seasonal weather type Local rain-and-clear dates and a contemporary shower-to-sunset report support its recurrence.
    • UnknownThe painted day No record identifies Heade’s observation or the exact marsh viewpoint.

What this cannot tell you. The clear-sky count is a lower bound from a single-writer transcription, not sunshine hours or measured shower duration. Other rain dates may also have had sunshine. The local observer’s sky is not the scene itself. Neither wind direction, tide stage nor a precise season date can be assigned to the canvas. The modern normals have standard S flags and may include NOAA estimates of missing months. No station pool, historical anomaly or climate trend is inferred.

Step 1 · Read the picture

A shower crossing the hay meadow

Nephos’s reading of the picture, not a museum statement: Salt-marsh hay stacked on the meadow, a dark shower cloud with visible rain shafts falling at left while sunbeams break through and the foreground marsh is lit: a passing summer shower during the salt-hay season.

The piled hay and small harvesting figures establish the work below; the dark left-hand cloud and vertical rain curtains establish precipitation in the distance. Diagonal shafts of light and bright bands across the meadow establish sunshine at the same pictured moment. We read the juxtaposition as passing weather. It does not reveal the storm’s direction of travel or a particular day.

The Met independently describes Dark rainclouds give way to sunrays and identifies a salt-hay harvest in this tidal landscape. The test is therefore specific: did rain and a clear daytime sky recur during the local hay season?

Mass Humanities places cutting and stacking on the Newbury salt meadows from late June through September. We use the full months July–September as a broad season, not a painted month. Contemporary evidence anchors its late-summer work: on 2 August 1881 the local paper reported a two-week interruption; three days later it reported hay being stacked.

“The wet weather and high tides of the past two weeks have prevented the Oldtown farmers from making their salt hay on the marshes.”

Newburyport Daily Herald, 2 August 1881, p. 1, “Oldtown. Salt Hay and Onions.” The report names both wet weather and high tides; it does not attribute the interruption to showers alone.

The Met dates the canvas about 1876–81. Its 1986 and 1987 publications proposed about 1872–78 from the first owner’s history. Neither supplies an observation day, and the accessible sources do not explain the revision. The rain sample below is selected for completeness, not resemblance to Heade’s sky.

Step 2 · Test the weather type

Rain did not always take the whole day

At least 5 of 48 rain dates also have a clear daytime sky entry. The two complete July–September seasons contain 184 daily rain records: 25 wet dates in 1879 and 23 in 1880. The confirmed minimum is 3 dates in 1879 and 2 in 1880. That is positive evidence for recurring changes between rain and daylight in the harvest season, rather than a search for Heade’s particular storm.

Karl Castelhun’s digitized local register supplies rain amounts. His original Signal Service Form H sheets add upper and lower cloud observations at 7 a.m., 2 p.m. and 9 p.m. The printed key defines zero as entirely clear. We accept a date only when it records positive rain and explicit zeros in both cloud layers at 7 a.m. or 2 p.m. The night column is excluded.

Rain and returning clear sky in the Newburyport hay seasonTwo July–September calendars contain 48 dates with recorded rain. At least 5 also have explicit zero upper and lower cloud amounts at 7 a.m. or 2 p.m. Dark squares mark rain; red squares with white circles mark rain plus a clear daytime observation. Open squares are zero rain, with sky not tested. The table below gives the observations.151015202530Rain / clear sky*1879Jul1 July 1879: 0.00 inches; sky not screened2 July 1879: 0.00 inches; sky not screened3 July 1879: 0.00 inches; sky not screened4 July 1879: 0.00 inches; sky not screened5 July 1879: 0.00 inches; sky not screened6 July 1879: 0.00 inches; sky not screened7 July 1879: 0.29 inches; clear daytime sky not established8 July 1879: 0.33 inches; clear daytime sky not established9 July 1879: 0.00 inches; sky not screened10 July 1879: 0.00 inches; sky not screened11 July 1879: 0.00 inches; sky not screened12 July 1879: 0.22 inches; explicit clear sky at 14:0013 July 1879: 0.00 inches; sky not screened14 July 1879: 0.00 inches; sky not screened15 July 1879: 0.00 inches; sky not screened16 July 1879: 0.19 inches; explicit clear sky at 07:0017 July 1879: 0.00 inches; sky not screened18 July 1879: 0.08 inches; clear daytime sky not established19 July 1879: 0.00 inches; sky not screened20 July 1879: 0.00 inches; sky not screened21 July 1879: 0.00 inches; sky not screened22 July 1879: 0.14 inches; clear daytime sky not established23 July 1879: 0.04 inches; clear daytime sky not established24 July 1879: 0.00 inches; sky not screened25 July 1879: 0.00 inches; sky not screened26 July 1879: 1.84 inches; clear daytime sky not established27 July 1879: 0.00 inches; sky not screened28 July 1879: 0.00 inches; sky not screened29 July 1879: 0.03 inches; clear daytime sky not established30 July 1879: 0.00 inches; sky not screened31 July 1879: 0.00 inches; sky not screened9 / ≥2Aug1 August 1879: 0.00 inches; sky not screened2 August 1879: 0.00 inches; sky not screened3 August 1879: 0.00 inches; sky not screened4 August 1879: 0.00 inches; sky not screened5 August 1879: 0.00 inches; sky not screened6 August 1879: 0.00 inches; sky not screened7 August 1879: 0.00 inches; sky not screened8 August 1879: 0.02 inches; clear daytime sky not established9 August 1879: 0.04 inches; clear daytime sky not established10 August 1879: 0.00 inches; sky not screened11 August 1879: 0.00 inches; sky not screened12 August 1879: 0.00 inches; sky not screened13 August 1879: 0.00 inches; sky not screened14 August 1879: 0.22 inches; clear daytime sky not established15 August 1879: 0.00 inches; sky not screened16 August 1879: 0.00 inches; sky not screened17 August 1879: 0.00 inches; sky not screened18 August 1879: 4.20 inches; clear daytime sky not established19 August 1879: 0.03 inches; clear daytime sky not established20 August 1879: 0.00 inches; sky not screened21 August 1879: 0.00 inches; sky not screened22 August 1879: 0.00 inches; sky not screened23 August 1879: 0.00 inches; sky not screened24 August 1879: 0.00 inches; sky not screened25 August 1879: 0.01 inches; clear daytime sky not established26 August 1879: 0.00 inches; sky not screened27 August 1879: 0.00 inches; sky not screened28 August 1879: 0.00 inches; sky not screened29 August 1879: 0.00 inches; sky not screened30 August 1879: 0.15 inches; clear daytime sky not established31 August 1879: 0.00 inches; sky not screened7 / ≥0Sep1 September 1879: 0.00 inches; sky not screened2 September 1879: 0.10 inches; clear daytime sky not established3 September 1879: 0.00 inches; sky not screened4 September 1879: 0.62 inches; clear daytime sky not established5 September 1879: 0.00 inches; sky not screened6 September 1879: 0.00 inches; sky not screened7 September 1879: 0.00 inches; sky not screened8 September 1879: 0.09 inches; clear daytime sky not established9 September 1879: 0.30 inches; clear daytime sky not established10 September 1879: 0.00 inches; sky not screened11 September 1879: 0.03 inches; explicit clear sky at 07:0012 September 1879: 0.00 inches; sky not screened13 September 1879: 0.00 inches; sky not screened14 September 1879: 0.29 inches; clear daytime sky not established15 September 1879: 0.00 inches; sky not screened16 September 1879: 0.00 inches; sky not screened17 September 1879: 0.08 inches; clear daytime sky not established18 September 1879: 0.00 inches; sky not screened19 September 1879: 0.00 inches; sky not screened20 September 1879: 0.00 inches; sky not screened21 September 1879: 0.00 inches; sky not screened22 September 1879: 0.07 inches; clear daytime sky not established23 September 1879: 0.00 inches; sky not screened24 September 1879: 0.33 inches; clear daytime sky not established25 September 1879: 0.00 inches; sky not screened26 September 1879: 0.00 inches; sky not screened27 September 1879: 0.00 inches; sky not screened28 September 1879: 0.00 inches; sky not screened29 September 1879: 0.00 inches; sky not screened30 September 1879: 0.00 inches; sky not screened9 / ≥125 rain dates; at least 3 with clear daytime sky1880Jul1 July 1880: 0.00 inches; sky not screened2 July 1880: 0.48 inches; clear daytime sky not established3 July 1880: 0.62 inches; clear daytime sky not established4 July 1880: 0.00 inches; sky not screened5 July 1880: 0.09 inches; clear daytime sky not established6 July 1880: 0.33 inches; clear daytime sky not established7 July 1880: 0.00 inches; sky not screened8 July 1880: 0.00 inches; sky not screened9 July 1880: 0.00 inches; sky not screened10 July 1880: 0.00 inches; sky not screened11 July 1880: 0.00 inches; sky not screened12 July 1880: 0.00 inches; sky not screened13 July 1880: 0.92 inches; clear daytime sky not established14 July 1880: 0.00 inches; sky not screened15 July 1880: 0.11 inches; clear daytime sky not established16 July 1880: 0.02 inches; clear daytime sky not established17 July 1880: 0.21 inches; clear daytime sky not established18 July 1880: 0.00 inches; sky not screened19 July 1880: 0.00 inches; sky not screened20 July 1880: 1.52 inches; clear daytime sky not established21 July 1880: 0.08 inches; clear daytime sky not established22 July 1880: 0.00 inches; sky not screened23 July 1880: 0.86 inches; clear daytime sky not established24 July 1880: 0.00 inches; sky not screened25 July 1880: 0.00 inches; sky not screened26 July 1880: 0.00 inches; sky not screened27 July 1880: 0.00 inches; sky not screened28 July 1880: 0.00 inches; sky not screened29 July 1880: 0.03 inches; explicit clear sky at 07:0030 July 1880: 0.00 inches; sky not screened31 July 1880: 0.00 inches; sky not screened12 / ≥1Aug1 August 1880: 0.00 inches; sky not screened2 August 1880: 0.00 inches; sky not screened3 August 1880: 0.00 inches; sky not screened4 August 1880: 1.31 inches; clear daytime sky not established5 August 1880: 0.01 inches; clear daytime sky not established6 August 1880: 0.00 inches; sky not screened7 August 1880: 0.00 inches; sky not screened8 August 1880: 0.00 inches; sky not screened9 August 1880: 0.00 inches; sky not screened10 August 1880: 0.15 inches; explicit clear sky at 07:0011 August 1880: 0.00 inches; sky not screened12 August 1880: 0.00 inches; sky not screened13 August 1880: 0.00 inches; sky not screened14 August 1880: 0.00 inches; sky not screened15 August 1880: 0.00 inches; sky not screened16 August 1880: 0.00 inches; sky not screened17 August 1880: 0.00 inches; sky not screened18 August 1880: 0.00 inches; sky not screened19 August 1880: 0.00 inches; sky not screened20 August 1880: 0.00 inches; sky not screened21 August 1880: 0.00 inches; sky not screened22 August 1880: 0.00 inches; sky not screened23 August 1880: 0.00 inches; sky not screened24 August 1880: 0.00 inches; sky not screened25 August 1880: 0.05 inches; clear daytime sky not established26 August 1880: 0.00 inches; sky not screened27 August 1880: 0.00 inches; sky not screened28 August 1880: 0.00 inches; sky not screened29 August 1880: 0.00 inches; sky not screened30 August 1880: 0.66 inches; clear daytime sky not established31 August 1880: 0.00 inches; sky not screened5 / ≥1Sep1 September 1880: 0.00 inches; sky not screened2 September 1880: 0.00 inches; sky not screened3 September 1880: 0.00 inches; sky not screened4 September 1880: 0.00 inches; sky not screened5 September 1880: 0.00 inches; sky not screened6 September 1880: 0.00 inches; sky not screened7 September 1880: 0.00 inches; sky not screened8 September 1880: 0.00 inches; sky not screened9 September 1880: 0.26 inches; clear daytime sky not established10 September 1880: 0.39 inches; clear daytime sky not established11 September 1880: 0.00 inches; sky not screened12 September 1880: 0.00 inches; sky not screened13 September 1880: 0.00 inches; sky not screened14 September 1880: 0.73 inches; clear daytime sky not established15 September 1880: 0.25 inches; clear daytime sky not established16 September 1880: 0.00 inches; sky not screened17 September 1880: 0.00 inches; sky not screened18 September 1880: 0.04 inches; clear daytime sky not established19 September 1880: 0.00 inches; sky not screened20 September 1880: 0.00 inches; sky not screened21 September 1880: 0.00 inches; sky not screened22 September 1880: 0.00 inches; sky not screened23 September 1880: 0.00 inches; sky not screened24 September 1880: 0.00 inches; sky not screened25 September 1880: 0.00 inches; sky not screened26 September 1880: 0.00 inches; sky not screened27 September 1880: 0.15 inches; clear daytime sky not established28 September 1880: 0.00 inches; sky not screened29 September 1880: 0.00 inches; sky not screened30 September 1880: 0.00 inches; sky not screened6 / ≥023 rain dates; at least 2 with clear daytime skyRecorded rainRain + clear daytime skyZero rain entered*Minimum confirmed from the scans. Other rain dates may also have had sunshine.
Newburyport, Karl Castelhun, station 195287. Rain: NOAA/CDMP and MRCC; clouds: Nephos transcription of the original NOAA NCEI scans. Every mark is computed from hash-checked inputs. A clear observation on a rain date is compatible with sun before or after rain; it does not measure a shower’s duration or establish simultaneous rain and sunshine over the meadow.

This is deliberately a lower bound. Sun can break through a partly cloudy sky, just as it does in the painting. Such entries do not pass the stricter fully-clear rule; neither do faint or blank cells. The other 43 rain dates are unconfirmed by this transcription, not classified as sunless. Two daytime observations can also miss a brief opening altogether.

One timed example: 29 July 1880. The original sheet records clear sky at 07:00 and rain from 3 pm to 4 pm: a 60-minute rain interval after a clear morning. This directly establishes a passing rain episode during one hay-season day; it is not proposed as Heade’s day.

“A beautiful sunset after the fine showers of Sunday.”

Newburyport Daily Herald, 26 August 1878, p. 3, local notices. This names the sequence directly for the preceding Sunday, 25 August. It is an additional seasonal witness, outside the two complete gauge seasons, and is not added to their counts.

The observer was in Newburyport; the gauge’s position is estimated and the painted viewpoint is unlocated. His sky is a local comparison, not the scene itself. The archive’s daily rain column has an unknown observation hour, so the join retains its dated rows rather than reconstructing hourly rainfall. No Boston or Newbury readings fill gaps.

Every daily amount and accepted clear-sky observation
DateRain, inchesDaytime clear-sky test
1 July 18790.00Not screened
2 July 18790.00Not screened
3 July 18790.00Not screened
4 July 18790.00Not screened
5 July 18790.00Not screened
6 July 18790.00Not screened
7 July 18790.29Not established
8 July 18790.33Not established
9 July 18790.00Not screened
10 July 18790.00Not screened
11 July 18790.00Not screened
12 July 18790.22Clear at 14:00; upper 0, lower 0
13 July 18790.00Not screened
14 July 18790.00Not screened
15 July 18790.00Not screened
16 July 18790.19Clear at 07:00; upper 0, lower 0
17 July 18790.00Not screened
18 July 18790.08Not established
19 July 18790.00Not screened
20 July 18790.00Not screened
21 July 18790.00Not screened
22 July 18790.14Not established
23 July 18790.04Not established
24 July 18790.00Not screened
25 July 18790.00Not screened
26 July 18791.84Not established
27 July 18790.00Not screened
28 July 18790.00Not screened
29 July 18790.03Not established
30 July 18790.00Not screened
31 July 18790.00Not screened
1 August 18790.00Not screened
2 August 18790.00Not screened
3 August 18790.00Not screened
4 August 18790.00Not screened
5 August 18790.00Not screened
6 August 18790.00Not screened
7 August 18790.00Not screened
8 August 18790.02Not established
9 August 18790.04Not established
10 August 18790.00Not screened
11 August 18790.00Not screened
12 August 18790.00Not screened
13 August 18790.00Not screened
14 August 18790.22Not established
15 August 18790.00Not screened
16 August 18790.00Not screened
17 August 18790.00Not screened
18 August 18794.20Not established
19 August 18790.03Not established
20 August 18790.00Not screened
21 August 18790.00Not screened
22 August 18790.00Not screened
23 August 18790.00Not screened
24 August 18790.00Not screened
25 August 18790.01Not established
26 August 18790.00Not screened
27 August 18790.00Not screened
28 August 18790.00Not screened
29 August 18790.00Not screened
30 August 18790.15Not established
31 August 18790.00Not screened
1 September 18790.00Not screened
2 September 18790.10Not established
3 September 18790.00Not screened
4 September 18790.62Not established
5 September 18790.00Not screened
6 September 18790.00Not screened
7 September 18790.00Not screened
8 September 18790.09Not established
9 September 18790.30Not established
10 September 18790.00Not screened
11 September 18790.03Clear at 07:00; upper 0, lower 0
12 September 18790.00Not screened
13 September 18790.00Not screened
14 September 18790.29Not established
15 September 18790.00Not screened
16 September 18790.00Not screened
17 September 18790.08Not established
18 September 18790.00Not screened
19 September 18790.00Not screened
20 September 18790.00Not screened
21 September 18790.00Not screened
22 September 18790.07Not established
23 September 18790.00Not screened
24 September 18790.33Not established
25 September 18790.00Not screened
26 September 18790.00Not screened
27 September 18790.00Not screened
28 September 18790.00Not screened
29 September 18790.00Not screened
30 September 18790.00Not screened
1 July 18800.00Not screened
2 July 18800.48Not established
3 July 18800.62Not established
4 July 18800.00Not screened
5 July 18800.09Not established
6 July 18800.33Not established
7 July 18800.00Not screened
8 July 18800.00Not screened
9 July 18800.00Not screened
10 July 18800.00Not screened
11 July 18800.00Not screened
12 July 18800.00Not screened
13 July 18800.92Not established
14 July 18800.00Not screened
15 July 18800.11Not established
16 July 18800.02Not established
17 July 18800.21Not established
18 July 18800.00Not screened
19 July 18800.00Not screened
20 July 18801.52Not established
21 July 18800.08Not established
22 July 18800.00Not screened
23 July 18800.86Not established
24 July 18800.00Not screened
25 July 18800.00Not screened
26 July 18800.00Not screened
27 July 18800.00Not screened
28 July 18800.00Not screened
29 July 18800.03Clear at 07:00; upper 0, lower 0
30 July 18800.00Not screened
31 July 18800.00Not screened
1 August 18800.00Not screened
2 August 18800.00Not screened
3 August 18800.00Not screened
4 August 18801.31Not established
5 August 18800.01Not established
6 August 18800.00Not screened
7 August 18800.00Not screened
8 August 18800.00Not screened
9 August 18800.00Not screened
10 August 18800.15Clear at 07:00; upper 0, lower 0
11 August 18800.00Not screened
12 August 18800.00Not screened
13 August 18800.00Not screened
14 August 18800.00Not screened
15 August 18800.00Not screened
16 August 18800.00Not screened
17 August 18800.00Not screened
18 August 18800.00Not screened
19 August 18800.00Not screened
20 August 18800.00Not screened
21 August 18800.00Not screened
22 August 18800.00Not screened
23 August 18800.00Not screened
24 August 18800.00Not screened
25 August 18800.05Not established
26 August 18800.00Not screened
27 August 18800.00Not screened
28 August 18800.00Not screened
29 August 18800.00Not screened
30 August 18800.66Not established
31 August 18800.00Not screened
1 September 18800.00Not screened
2 September 18800.00Not screened
3 September 18800.00Not screened
4 September 18800.00Not screened
5 September 18800.00Not screened
6 September 18800.00Not screened
7 September 18800.00Not screened
8 September 18800.00Not screened
9 September 18800.26Not established
10 September 18800.39Not established
11 September 18800.00Not screened
12 September 18800.00Not screened
13 September 18800.00Not screened
14 September 18800.73Not established
15 September 18800.25Not established
16 September 18800.00Not screened
17 September 18800.00Not screened
18 September 18800.04Not established
19 September 18800.00Not screened
20 September 18800.00Not screened
21 September 18800.00Not screened
22 September 18800.00Not screened
23 September 18800.00Not screened
24 September 18800.00Not screened
25 September 18800.00Not screened
26 September 18800.00Not screened
27 September 18800.15Not established
28 September 18800.00Not screened
29 September 18800.00Not screened
30 September 18800.00Not screened
Coverage across both published date ranges
YearUsable daily rain entriesUse
18720 / 92Incomplete; no seasonal result
187314 / 92Incomplete; no seasonal result
18740 / 92Incomplete; no seasonal result
18750 / 92Incomplete; no seasonal result
18760 / 92Incomplete; no seasonal result
18770 / 92Incomplete; no seasonal result
18780 / 92Incomplete; no seasonal result
187992 / 92Complete; tested
188092 / 92Complete; tested
18810 / 92Incomplete; no seasonal result

Only 1879 and 1880 have complete, unflagged local rain seasons in this file. Their survival does not favour the current catalogue date over the earlier one. Cloud readings are a single-writer transcription of faint historical sheets; an independent transcription check remains desirable.

Step 3 · Then and now

The hay season still carries rain

For the same three months, NOAA’s 1991–2020 Newburyport normals add to 28.7 days with at least 0.01 inch of rain and 10.90 inches. The historical totals are 9.74 and 9.27 inches. All positive historical amounts meet that same rain-day threshold.

These seasons and the modern normal all describe a harvest period repeatedly exposed to rain. The modern numbers belong to a different Newburyport station, USC00195285. They do not measure how often showers brought sunshine, and two historical seasons cannot establish a climate trend.

The same July–September season; station records kept separate
PeriodRecordRain, inchesDays ≥0.01 inchRain + clear daytime sky
July–September 1879Castelhun, 1952879.7425≥3
July–September 1880Castelhun, 1952879.2723≥2
1991–2020 averageNewburyport, USC0019528510.9028.7Not tested
Monthly normals and their completeness
MonthRain, inchesDays ≥0.01 inchNOAA flag / available years
July3.5110.2S; 28 years
August3.499.4S; 30 years
September3.909.1S; 30 years

All three monthly normals carry NOAA’s standard “S” completeness flag. NOAA’s method can estimate missing months using surrounding stations. We sum the published monthly normals at this one modern station; we do not pool its observations with Castelhun’s. Historical gauge exposure and continuity with the modern site are unestablished, so no percentage change or anomaly is claimed.

Step 4 · What Heade chose

The instant of release, with work still under way

Nephos reads the painting’s decision as the coexistence of rain and light: the shower remains visible while the meadow is illuminated. The workers and haystacks make that passing weather consequential. The historical records establish that rain and clear daylight could share a hay-season date; the newspaper shows why the intervals mattered to work on the marsh.

The Met’s 1987 catalogue describes a deliberate reduction of foreground detail to pull attention into the distance. It situates the canvas among less rigid marsh compositions, with a lower, broader horizon and less regular haystacks. That is a supported account of arrangement, not proof that Heade moved a particular stack or repainted this storm.

Elizabeth Leto Fulton’s technical study of Heade’s paintings found that he generally developed compositions on the canvas, beginning with a simple drawing and adjusting forms in paint. Her examination identified changes to motifs including trees and haystacks. This is evidence about his working practice across other paintings, not a technical examination of Met 1985.117. It strengthens the case for treating the scene as a considered composition without inventing its making history.

The earlier Met catalogue cautiously states He seems not to have relied on preparatory sketches in its discussion of marsh scenes. That does not mean no Heade drawings survive. No dated sketch, letter or outdoor session has been tied to this canvas here. The artist’s stretcher inscription Newburyport Meadows, transcribed in the acquisition account, names the locality.

A useful caution comes from the harvest itself. On 5 August 1881 the Herald reported hay being stacked, including much of it left over from last season has been blackened by rain and tides. Stacks could include older material. Their presence alone cannot date the picture; the harvesting figures and the broader seasonal evidence support our reading of active haying.

What remains unresolved: the painted day and exact viewpoint; whether the shower is approaching or leaving; its duration, rain amount and tide stage; and whether Heade changed weather he had observed. The supported finding is narrower and useful: the picture’s rain-and-light conjunction belongs to a documented local harvest environment, and Heade made that change of light the centre of the composition.

Sources

Every source on this page

Each file was downloaded and given a fingerprint (its sha256). The page is rebuilt from these files and will not build if a fingerprint changes, so every number above comes from them rather than being typed in.

  1. The Met Open Access API, object 11053CC0 object metadata · retrieved 7 October 2026 · 1985.117; display date and numeric bounds kept separatelysha256 fb689905ee12bbb53810e5bf9178a594f4e37bde6fdaac64ae00a01b85d13a38
  2. The Met, Newburyport Meadows, current labelCopyright Metropolitan Museum of Art; short quotation and research cache · retrieved 7 October 2026 · Web-tool text capture; direct HTTP download returned 429sha256 4007861f162245215a8cc29ce9489ea7a5fba619bebab080e7642da6da2ef59e
  3. The Met, Recent Acquisitions: A Selection, 1985–1986Copyright 1986 Metropolitan Museum of Art; publisher provides free research PDF · retrieved 7 October 2026 · p. 51 / PDF p. 52; rendered scan inspectedsha256 b0f2d94cd4b2a9ff779742017c73da524e064c00b204d04fbc1a359944e9074d
  4. Short excerpts transcribed from the acquisition accountShort research transcription of copyrighted text · retrieved 7 October 2026 · Nephos transcription, line breaks normalized; not publisher OCRsha256 cb2cb8b7aefa14f0483f944cdde6a333c777f948ce719b52a44bfef43df94986
  5. The Met, American Paradise (1987), Newburyport MeadowsCopyright 1987 Metropolitan Museum of Art; library provides research PDF · retrieved 7 October 2026 · p. 177; rendered scan inspectedsha256 470a31046d7d1bdcd8d273418a64536dc57eb9b5953def5ef1ee4a8e693deafb
  6. The Met, American Paradise (1987), making and provenanceCopyright 1987 Metropolitan Museum of Art; library provides research PDF · retrieved 7 October 2026 · p. 178; rendered scan inspectedsha256 d2922fa4472ab20505acbfbb6a5725f00b292107877f9082bed8da3e3fbaaca8
  7. Text extraction of American Paradise, p. 177Copyright 1987 Metropolitan Museum of Art; research extraction · retrieved 7 October 2026 · PyMuPDF extraction of cached PDF; selected quotations checked on scansha256 bb63bc0a2bc4cafbb407af674a2ccc0b383ce1c95ff10dbfc3c403a8d8b79c99
  8. Text extraction of American Paradise, p. 178Copyright 1987 Metropolitan Museum of Art; research extraction · retrieved 7 October 2026 · PyMuPDF extraction of cached PDF; selected quotations checked on scansha256 805d123d206a88f1ae4fd3c494e119ef0dfbd34ac85049f9a92bcd40ce09e9f2
  9. Newburyport daily observations, NOAA/CDMP, keyed and checked by MRCCHistorical NOAA/CDMP observations; credit MRCC · retrieved 7 October 2026 · Station 195287 only; PRCP-99, inches; Newbury station 195282 remains separatesha256 29744a9d1f26366a43ed77a86ea54d6a422173e60c4221bde3bf56d128985c5b
  10. MRCC, Newburyport and Newbury station historiesHistorical NOAA/CDMP metadata; credit MRCC · retrieved 7 October 2026 · Karl Castelhun, May 1879–February 1881; estimated positionsha256 760405af13559b45081c1e30a315019b8ab39b4db8727b7db9dad18dfe3526c7
  11. NOAA/MRCC, CDMP Forts format, version 1.1 (March 2010)US government documentation · retrieved 7 October 2026 · Units, trace and quality flags; no cloud or sunshine field in the selected CSVsha256 3ade954c6c3f982ce8ab390f776f17cee7dce1b884e33e955a32a780db5a7cdf
  12. Mass Humanities, Mass Moments: Newbury Offers Wolf BountyCopyright Mass Humanities; short quotation and research cache · retrieved 7 October 2026 · Historical account of salt-hay cutting, late June–September; not a date for the canvassha256 013604fa00576904f25a2a7114c10f5e0f6eb3eb67e3561a5f6f4274b476da81
  13. Martin Johnson Heade, Newburyport Meadows, Met 1985.117Public domain / Met Open Access; isPublicDomain true in cached object API · retrieved 7 October 2026 · Full composition; only resizing and JPEG encoding for displaysha256 25d6e075c5424864c86df672b0f5789f081e0dc9a37ac1f2c4ee196e0d881a16
  14. Karl Castelhun, Newburyport Form H, 1879-07Public-domain US Signal Service record; NOAA NCEI scan · retrieved 8 October 2026 · CL-734 / AE04529 / image 00144; daylight cloud columns, 7 a.m. and 2 p.m.; 9 p.m. excludedsha256 591e31c351a8e5d46e83b9336ff90dfe7217af69c7c6549b7743ece1cababfb0
  15. Karl Castelhun, Newburyport Form H, 1879-08Public-domain US Signal Service record; NOAA NCEI scan · retrieved 8 October 2026 · CL-734 / AE04529 / image 00146; daylight cloud columns, 7 a.m. and 2 p.m.; 9 p.m. excludedsha256 f520da1a1fb576b8270ebe34ecc45d32a8ed06d19678dc0703849e596d916a4a
  16. Karl Castelhun, Newburyport Form H, 1879-09Public-domain US Signal Service record; NOAA NCEI scan · retrieved 8 October 2026 · CL-734 / AE04529 / image 00150; daylight cloud columns, 7 a.m. and 2 p.m.; 9 p.m. excludedsha256 425e6247f95eb62fb46c311611edbcc137747127e01973db38241ab2a61aa5cd
  17. Karl Castelhun, Newburyport Form H, 1880-07Public-domain US Signal Service record; NOAA NCEI scan · retrieved 8 October 2026 · CL-734 / AE04529 / image 00172; daylight cloud columns, 7 a.m. and 2 p.m.; 9 p.m. excludedsha256 50addf274331b4d73b67bd6dfc2e5d21c4602eb982b9afc53869fc4cfc6d5fb4
  18. Karl Castelhun, Newburyport Form H, 1880-08Public-domain US Signal Service record; NOAA NCEI scan · retrieved 8 October 2026 · CL-734 / AE04529 / image 00174; daylight cloud columns, 7 a.m. and 2 p.m.; 9 p.m. excludedsha256 03715a764dbd834a0192d446c65af1f9ba2e439f3f320f61895580686881ca80
  19. Karl Castelhun, Newburyport Form H, 1880-09Public-domain US Signal Service record; NOAA NCEI scan · retrieved 8 October 2026 · CL-734 / AE04529 / image 00176; daylight cloud columns, 7 a.m. and 2 p.m.; 9 p.m. excludedsha256 8e394876aa90bc4fd02845dcab8e87224cafe59e11b62c4a77e0ce8c95498db2
  20. Newburyport Daily Herald 18780826, p. 3US newspaper published before 1931; public domain · retrieved 8 October 2026 · Newburyport Daily Herald 18780826, p. 3; page image checked, OCR used only for locating textsha256 1c05acc057ec8f455fccfe706f5429faec3c200662d78885807caf6caacf07f4
  21. Newburyport Daily Herald 18810802, p. 1US newspaper published before 1931; public domain · retrieved 8 October 2026 · Newburyport Daily Herald 18810802, p. 1; page image checked, OCR used only for locating textsha256 006b7083a81d00f4a98c38a9a6dcf1722ff534f880f11b01a1da9cc8c5435baa
  22. Newburyport Daily Herald 18810805, p. 1US newspaper published before 1931; public domain · retrieved 8 October 2026 · Newburyport Daily Herald 18810805, p. 1; page image checked, OCR used only for locating textsha256 408758eac4379d600866390e4b89cd74d319f3332d7375894b1c004e4e01fb32
  23. Nephos transcription: explicit daytime clear skies on rain datesNephos transcription of public-domain observations · retrieved 8 October 2026 · Five accepted zero-pairs; identical committed evidence file; non-accepted dates remain unconfirmedsha256 b641917d187d03e8cd3fd50b8128c031faa42cb5636055a4b342f972d5426e21
  24. Nephos transcriptions of three Newburyport Herald witnessesNephos transcription of public-domain newspaper text · retrieved 8 October 2026 · Source-exact excerpts checked on the page scans; normalized line breaks onlysha256 7cc61d21739074d93f8b63a626d7a92cf592173d745cdc4d3f21314d9ddd8a3e
  25. NOAA NCEI, Newburyport 1991–2020 monthly normalsNOAA public data · retrieved 8 October 2026 · USC00195285; MLY-PRCP-NORMAL and MLY-PRCP-AVGNDS-GE001HI; July–September; completeness flags Ssha256 bff59bed1976c8104446dfd4d7208a740fa7d20caa17fa386e214af86508ffaf
  26. NOAA NCEI, monthly normals documentationUS government documentation · retrieved 8 October 2026 · 1991–2020 monthly normals; Table 1 completeness flags, Table 2 rainfall-day threshold >=0.01 inchessha256 b12c9aee6770f84b5fa902740413bffca5cbf9a83bbebcf7e73306f46168ffc6
  27. Elizabeth Leto Fulton, The Methods and Materials of Martin Johnson Heade (JAIC, 2002)JAIC/AIC copyrighted text; publicly available research copy; brief attributed paraphrase only · retrieved 8 October 2026 · Sections 2 and 6: composition developed and adjusted on canvas; general technical study, not examination of Met 1985.117sha256 785787f092a2b3ecdade7f0683527d888c724fe691d0091aa925f307835b32fe

The numbers on this page, as JSONThe method card, as JSON