A crop year does not feel like recordkeeping while it is happening. It feels like weather, and equipment, and a hundred decisions made standing up. You plan a field, plant it, walk it, spray or feed or amend it, and something comes off. The paperwork is the part you do afterward — the part nobody trained you on.
But the season is not really a sequence of independent jobs. It is a chain — crop plan, planting, scouting, input application, harvest lot — and its value is not the strength of any single link. It is that the links connect. A harvest lot that can name every input that ever touched it is worth something; the same lot with the connections missing is a weight and a date. This page is about those connections: what each record has to carry so the next one can find it, and why the season’s paperwork is tested not the day you write it but months later.
The label and your certifier are the authority. This page does not tell you a re-entry interval, a pre-harvest interval, or any other regulated figure, and you should distrust anything that does. Those intervals come from the product label, your certifier, and your regulator; they vary by product, formulation, crop, and jurisdiction, and they change. When this page uses a number it is visibly hypothetical. What follows is about the structure of the record — how the links connect — not about what any interval is. For the interval, read the label and ask the person whose job it is to know.
The value of a crop year lives in the links, not the nodes
Almost every farm can produce the nodes. Ask what was planted and you get a list; what was sprayed, a spray log; what came off, a scale ticket somewhere. Each node exists because a moment forced it into being. The links are different: nothing in the day forces you to connect the spray log to the field to the harvest, so that connection is a separate act of discipline, invisible until the moment someone needs it — at which point its absence is the only thing anyone can see.
So think of a crop record not as a stack of documents but as a set of pointers. The planting points at a field and a seed lot. The input application points at a field and a window of dates. The harvest lot points back through the planting to everything ever addressed to the same ground. Break one pointer and a whole branch of the answer goes dark. This is the discipline under farm recordkeeping as a whole: the record you will need is the one whose connections you kept.
A planting record is written for a reader who does not exist yet
A planting record looks backward — it says what went into the ground — but its real job is forward. It is the anchor that a harvest lot, not yet imagined, will point back to in order to name every input that touched the crop. So it has to carry enough identity to be found and named later, by someone who was not there: the field or block, the crop and the specific variety, the seed or transplant source with its lot number, the date, and who did it. The seed lot is the field most often left blank and the one that matters earliest — the first link in the whole chain, the thing a germination or seed-borne problem reaches for before anything else.
A planting recorded as “corn, north field, May” is a memory. A variety, a seed lot, a block, and a date is an anchor. The test: could a stranger, reading only this line a year from now, say which crop, in which ground, from which seed, it names? If it depends on knowing the north field is the one past the shed, the record leans on a memory that will not be in the room when it matters.
Scouting is the record nobody keeps and everybody later wants
Of all the season’s records, scouting is the one with no moment that forces it — no seed to buy, no jug to open. Walking a field and seeing it looks fine forces nothing, so the observation evaporates the instant you climb back on the truck.
That is a real loss, because scouting answers the questions the other records cannot. Why did you apply that input, on that date, to that block? The input record says you did; the scouting note from three days earlier says why. Did a problem appear overnight or build for weeks? One note cannot say. Six weekly notes can. The value is not in any single observation but in the pattern it accumulates into, and a pattern exists only if the notes were captured while they were cheap: the block, the date, what you saw, and enough specificity that “pressure in the southeast corner” is a place a second person could go stand. The note you meant to write that evening does not count — by then the corner has blurred into the day.
Every input application is addressed to a place and a window
An input application record shares its skeleton with a livestock treatment record, for the same reason: a substance was applied, and time has to pass before the field is safe to enter or the crop safe to harvest. The fields that make it defensible — active ingredient, rate, the product’s lot number, the re-entry interval, the pre-harvest interval — are the whole subject of the spray and input records cluster, and they belong captured at the sprayer, not at the desk.
What matters at the level of the whole chain is what the application points at: a block, and a window of dates. Those two coordinates are how it will later join a harvest lot. If the record says “the tomatoes” and three tomato blocks were planted a month apart, it has lost the ability to say which harvest it constrains. The precision is not fussiness — it is the hinge the pre-harvest interval turns on. The same intervals reappear in livestock management as withdrawal periods, and the discipline is identical: the interval belongs to the label, and the record’s job is to attach it to the right thing so nothing inside an open window moves.
The harvest lot is the address everything downstream is written to
At some point in the season, product stops being “the crop in the field” and becomes a distinct, countable, sellable thing. That is the harvest lot, the most important unit in the crop record, because it is the address everything after harvest is written to. A customer does not buy your season; they buy a lot. A recall does not concern your farm; it concerns a lot. A buyer’s traceability request does not ask what you grew — it asks what a lot is and where it went.
Because so much points at it, the harvest lot has to point back: through the planting it came from to every input that touched the same block and window, and forward to every customer it shipped to. The lot is the join, where the two halves of traceability — one back to inputs, one up to customers — meet. So treat it as a first-class record with its own identity: a code, a crop, a source planting, a date, a quantity, a location. If it is thin, the whole chain terminates in something too vague to use, however careful the earlier links were.
A lot code assigned late is a lot code that means nothing
The most common way a harvest lot fails is not that it is missing but that it was invented after the fact, to fill a column a buyer’s form demanded. A code created a week after harvest, or the morning of a shipment, has a fatal defect: nothing was addressed to it while the things it is supposed to trace were actually happening.
Say, hypothetically, a block is harvested over three days, moved into two coolers, and drawn down across a month of sales. If the code exists from the first bin, every movement is written against it as it happens, and the lot at the end is a true account of a real thing. Assign it when the first customer asks, and someone has to reconstruct which cooler held which day’s pick and which sale drew from which — from memory, under pressure, which is how a lot code becomes a fiction with a plausible format. The code looks identical either way; only one is load-bearing. A lot code is an address you open before there is mail to deliver to it, not a label you apply to finished product — the distinction a food safety auditor is trained to find in about one question.
A field boundary is a data structure, not a map
Every record in the chain is addressed to a place, so the place has to be something records can agree on. When a field is only a name in someone’s head — the north field, the rented forty, the one past the creek — the records addressed to it are anchored to that person’s memory. The day they leave, or the day the field is split into two blocks, the anchor drifts, and records that once agreed quietly stop agreeing.
A field boundary held as data fixes this — not a pretty map, but a defined block with a stable identifier and a known area. The identifier lets a planting, an input application, and a harvest lot all say “this ground” and mean the identical ground. The area is what makes a rate or a total mean anything: a product applied at some rate per acre is meaningless until the acres are a fixed, agreed number. This is why splitting and merging fields is where crop records quietly break — the plantings and harvests addressed to the old shape have nowhere clean to live. A boundary that is versioned, identified, and carries its area is what lets the chain survive the farm changing shape, which every farm eventually does.
Joining the chain is where the work is, and where it fails
Everything above describes a chain that connects, and assembling it by hand — binders, scale tickets, a customer ledger — is how most farms do it. The chain is real whether or not software holds it; what software changes is only whether the joins are made for you, or by you at the moment someone asks.
Farm40 builds the traceability packet by joining a harvest lot to the inputs applied to it — linked through the shared planting — and forward to the customers it shipped to, so a request that meant an afternoon in the binders becomes a single assembled view. The limitation is the point, and it is the same one that governs every record on this page: the packet is only as complete as the data entered. A sale recorded with a blank lot code does not link to anything. The tool assembles the chain you kept; it cannot supply a link you skipped. If you would rather stay in a spreadsheet, the honest reading of farm software is that a disciplined spreadsheet with a real lot column beats an expensive system entered carelessly, every time.
The record’s real test comes months later, from someone who was not there
A crop record does not get graded the day you write it. It gets graded in the winter, or the following spring, or the afternoon a buyer’s auditor sits at your table — always later, and by someone who was not in the field. That reader has no memory to lean on; they have only what the record says, exactly as it says it.
This reverses the intuition about which records deserve care. The planting you record today feels important because the field is in front of you and the details are obvious — which is exactly why you are tempted to leave them thin. The record that will actually be read is the one from last season, whose field has been replanted and whose crew has turned over, and which therefore has to be complete on its own. So write every link for that reader: a planting that names its seed lot, a scouting note a stranger could stand in, a harvest lot that existed from the first bin. The season is the easy part; you already know how to farm. The chain that survives it is the part nobody trained you on, and it is built one honest link at a time.