The Harvesters

One panel of the most famous broken set in European painting. Bruegel painted a cycle of landscapes through the year in 1565 for the Antwerp merchant Niclaes Jongelinck; within a year the whole of Jongelinck's collection had been pledged to the city of Antwerp as security for somebody else's tax debt, and the document that records the pledge is the earliest thing anyone knows about these pictures. Three of the survivors are in Vienna, one is in Prague and this one is in New York.

Its route there runs through an archduke, an emperor, a Napoleonic general and a Paris couturier. In 1816 it was sold in Paris as the work of an unnamed master of the old German school, dated 1546, for ninety francs and eleven centimes. In 1919 the Metropolitan Museum bought it in a market flattened by the war, under dirt and varnish, and found the signature when it cleaned it.

What it is

Oil on four planks of oak, laid with the grain running horizontally, joined and originally dowelled, measuring 116.5 by 159.5 centimetres of original painted surface. A wheat field in the heat of the day, cut on the left, standing on the right, with a path driven through it. Under a pear tree in the middle a group has stopped to eat; one man lies on his back with his legs apart and his hat falling off, asleep or drunk. Beyond them the land drops away to a village, a church, an estuary and ships.

There is no story and no saint. That is the point of the picture and it is why it is usually called the beginning of European landscape painting as a subject in its own right. What Bruegel put in the middle of the field is a jug of beer or water, cooling in the wheat, and a man with a scythe looking at it.

One thing in the picture cannot be explained. At the upper left there is a blue sphere. The Met's catalogue entry sets out why it is not the harvest moon, which comes near the autumn equinox and could not stand that high in the sky at noon, and why it is not a solar eclipse, since the disc is not black, there is no corona, the landscape is not darkened and nobody in the field has looked up. The museum's position is that it is an enigma and may simply be the sun veiled by the haze of a late-summer midday. Recording the enigma is more useful than resolving it.

Made for a merchant, pledged against a tax debt

The cycle was painted in 1565 for Niclaes Jongelinck, an Antwerp merchant who had also bought Frans Floris's Seven Liberal Arts and Labours of Hercules and a set of Seven Planets by the sculptor Jacques Jonghelinck. Several scholars have argued the four commissions were meant to work together as one programme in his suburban house.

What survives from that moment is not a contract. It is a document in the Antwerp city archives dated 21 February 1565 by the old calendar, which is 1566 by ours, recording that Jongelinck's whole art collection was pledged as collateral for tax arrears owed to the city by Daniel de Bruyne. It names sixteen pieces by Bruegel and singles out three: the Tower of Babel, the Carrying of the Cross and the Twelve Months.

So the earliest evidence for the greatest landscape cycle of the sixteenth century is a security instrument for somebody else's unpaid tax. The pictures were painted, signed, hung and encumbered inside twelve months. Every later dispute about the series turns on the two Dutch words in that document, de Tweelff maenden, the twelve months, and on whether they describe twelve panels or one subject.

Six or twelve, and how many are left

The question is not academic. If there were twelve panels, seven are missing. If there were six, one is.

Gustav Gluck argued in 1932 that the five known works belonged to an incomplete series of twelve, and changed his mind in 1936. Charles de Tolnay had proposed in 1934 that the cycle was six pictures, each covering two successive months, and most recent scholarship follows him: not more than six could ever have existed. Fritz Grossman reverted to twelve in 1955, observing that calendar series always begin in January even where the official year began at Easter.

Two documents decide it as far as anything does. The 1594 Antwerp record and Blasius Huetter's account-book entry of 5 July 1594 both say six paintings representing the twelve months. The posthumous inventory of Archduke Ernst in 1595 lists six panels of the twelve months of the year by Bruegel, as nos. 7 to 12. The inventory of Archduke Leopold Wilhelm in 1659 lists five. One panel is lost between those two dates and no other picture has ever surfaced that could belong to the series or be a copy of a lost member of it.

If the cycle was six, it is among the earliest calendar series to divide the year into six parts rather than twelve. Not the only one: Novotny pointed in 1948 to a contemporary Months by Marten van Cleve in which the months are paired, and Huetter recorded the archduke buying six pictures of the seasons from a Namur painter, Jean Sayve, in February 1594.

Which months this one is

Even the title is contested. Tolnay, starting the year atypically at December and January, made this panel August and September. Gluck, following the calendar of the Spanish Netherlands and starting at March and April, made it July and August. Robert Delevoy, Claessens and Rousseau, and the 1991 Vienna catalogue follow Gluck. A late sixteenth-century drawing after this composition ascribed to Pieter Balten, in the Kupferstichkabinett in Berlin, is inscribed Der Monat Juli.

Others reject the whole framework. Novotny, Genaille and Demus argue that this is not a clear-cut series of months at all but an extended cycle of the seasons, or in Genaille's phrase the characteristic moments of each season.

The drawing that supports Tolnay is a different one: Pieter Stevens's sheet in the Steiner collection in New York, published by Van Miegroet in 1986, comparable to The Gloomy Day and inscribed februarius and mert. Iain Buchanan followed Van Miegroet and Tolnay in 1990. The debate is now more than ninety years old and it turns on two inscribed drawings by other artists.

The missing panel, and a bold proposal about it

It is generally agreed that the missing picture showed March and April. Bruegel left a drawing of Spring in the Albertina in Vienna, inscribed by him De Lenten Meert April Meij, and the lost panel is usually assumed to have resembled it at least in subject. His second drawing for that series, Summer, in the Hamburg Kunsthalle, is a harvest scene with elements in common with this painting but a quite different composition; Hans Bol completed Autumn and Winter and Hieronymus Cock published the four as an engraved series of the Four Seasons in 1570.

Reindert Falkenburg made a different proposal in 2017, and the Met's own curator reports it and then declines it. Falkenburg suggests the lost painting may not be lost: that it is the Procession to Calvary in the Kunsthistorisches Museum in Vienna, which drifts in and out of the same inventories as the Seasons panels, and that it represents the pivotal event in the history of salvation which the cycle is otherwise built around.

Maryan Ainsworth, writing the Met's entry, calls that a bold leap open for discussion and prefers a panel based on the Albertina Spring drawing. Both positions are set out here because the disagreement is between named scholars in print, and arbitrating it is not this record's business.

There is a related trap for anyone reading old inventories. Grossman, Demus and Buchanan agree that the paintings of the seasons listed in inventories of the Brussels palace in 1659, in the 1660s to 1690s and in 1732 must be either copies destroyed in the palace fire of 1731 or unrelated pictures. They are not these panels.

Where the wood came from, and what it proves

Dendrochronological analysis of the four planks gives an earliest possible creation date of 1550 and a more plausible one from 1556 onwards, which is consistent with the painted date of 1565. Two of the four boards came from the same tree. The wood is from the Baltic and Polish region.

That last fact is not a curiosity. Antwerp painters worked on oak shipped down from the eastern Baltic through the Hanseatic trade, and the timber trade is frequently the strongest dating and localisation evidence a Netherlandish panel has. The same analysis on the frames of the Met's van Eyck diptych ties that oak to the same part of the eastern Baltic as a panel now in Williamstown. A supply chain running from a Polish forest to an Antwerp workshop is doing evidential work here that no amount of connoisseurship could do.

Cut down, planed, cradled and put back

The panel is not the size Bruegel made it, and it has been through more physical intervention than almost anything else the Met holds from the century.

Only the left edge survives as an original edge, identifiable by a narrow unpainted margin and a barbe, and the Met's conservators note that the barbe was probably formed by a channel-edge support rather than an engaged frame. The top, right and bottom were all trimmed, by somewhere between 1.3 and 1.9 centimetres each, judging from the other panels of the series. Half an inch off the top left the blue sphere so close to the edge that a frame shadow nearly hid it; the cropping at the right bisected a bending figure.

At some point before the picture reached New York the oak was planed down to five millimetres, a strip nineteen millimetres deep was added at the bottom and a cradle was fixed to the back. In 1998 the Met removed the cradle, repaired three joins and several splits, and fitted a strainer with a spring mechanism to let the wood move. It kept the old bottom strip and added strips averaging thirteen millimetres to the top and right, which is why the tombstone quotes two sets of dimensions and why the larger of them is not a measurement of anything Bruegel painted.

Watching him change his mind

Under the paint is an extensive underdrawing in a dry medium, probably black chalk, laid over an off-white ground and a lead white priming. Because the oil has grown more translucent with age, much of it can now be seen with the naked eye.

Bruegel drew the composition freely and fast, repeating contours two or three times as he hunted for a position, exaggerating the hunched shoulder of the labourer at lower left and scribbling the sprawled figure at the centre into place. He even planned where the birds would go in the sky. The lines break slightly where they cross the ridges of the priming, which proves he drew on top of the imprimatura rather than under it.

Two late changes carry the argument about what the picture means. The hat falling from the sleeping man's head was painted over his finished sleeve at a late stage and has grown transparent, so the sleeve now shows through it. And the dead limbs hanging over both the sleeper and the group eating their meal have no underdrawing at all; Bruegel added them in paint alone, late, directly above the figures.

Sloth, diligence, and a jug of beer in the wheat

Three readings of the picture are in play and the physical evidence bears on two of them.

Bertram Kaschek reads the man with the scythe as about to destroy the vessel cooling in the wheat, an apocalyptic shock breaking in on the people resting. The Met's entry counters him with the underdrawing: Bruegel adjusted the scythe-man's pose but was consistently definite about his line of sight, and took care not to pull the hat down over his eyes, so that he looks straight at the jug. A labourer eyeing the drink he is about to be allowed is not an angel of judgement. Two later versions by Pieter Bruegel the Younger, sold at Christie's in London in 1981 and 2012, leave the vessel out altogether, which suggests his son did not read it as load-bearing either.

Lyckle de Vries reads the picture as an opposition of sloth and diligence, supported by proverbs and by biblical warnings against laziness. Here the physical evidence supports the reading: the dead branches were added late, in paint, directly over the sleeping man and the eaters. A tree withers for want of water as a man is paralysed by idleness, and Bruegel put the branches in after the fact, as if fine-tuning the moral.

Falkenburg reads the whole cycle as the presence of God in the created world, to be found by a viewer who looks closely: the fallen apples in the right background, the partly barren pear tree in the foreground recalling the tree in Eden and the fruit of the Virgin's womb, the natural year running beneath the history of salvation. The Met's entry sets out all three and observes that they could have coexisted, offering the wealthy, educated visitors to Jongelinck's country house something to argue about over the meal. That is very likely what the pictures were for.

Ninety francs, and a hundred and three years

The Andreossy sale at Perignon in Paris on 11 March 1816 offered the panel as lot 1: by a master of the old German school, a landscape of the greatest detail showing the labours of the harvest, dated 1546. It made 90 francs 11 centimes.

Nothing in that catalogue line is correct except the subject. The nationality is wrong, the date is nineteen years out, and the painter, far from anonymous, is one of the most reproduced names of the previous century. This is what a cataloguing failure costs, expressed in money: a Bruegel from the imperial collection sold for less than the price of a good clock because the room did not know what it was looking at.

A hundred and three years later the Metropolitan Museum bought it from Paul Jean Cels for a price its own curators describe as extraordinarily low, in a market flattened by the First World War, with the picture still under dirt, discoloured varnish and clumsy restoration along the joins. Bryson Burroughs took the risk and was right. The signature and the date came up when the panel was cleaned.

Afterlife

Copies and versions exist and they are the standing hazard for anyone matching a photograph to this record. A late sixteenth-century drawing after the composition ascribed to Pieter Balten is in the Kupferstichkabinett in Berlin as no. 13220, titled Der Monat Juli. Two painted versions by Pieter Bruegel the Younger, closely based on this panel, passed through Christie's in London in 1981 and 2012. A small copy on copper by an anonymous Flemish painter belonged to the Met itself until 1978, when it was deaccessioned and sold at Parke-Bernet, having carried another attribution for many years.

The picture's fame in New York is a twentieth-century construction and one witness to it is unexpected. Lillian Ross, profiling Ernest Hemingway in the New Yorker of 13 May 1950, recorded his comments in front of this painting; he admired the geometric handling of the grain.

The Met does not lend it. The object record carries the line in capitals. A panel of four planed oak boards, trimmed on three sides, cradled and uncradled, is not a travelling picture, and the other survivors of the cycle in Vienna are subject to the same reasoning. The set has not been reassembled and, on present conservation policy at both museums, will not be.