Wheat, as the official record holds it

The AgricultureID entry for wheat is the canonical account: what the crop is, how it grows, what it needs, what it is affected by. This piece is not that, and is deliberately not a summary of it. It is about the sources — what the official record holds on wheat, and how differently each part of it is defined.

Ask nine sources about wheat and you get nine different objects.

The economics source holds a forecast

USDA's Economic Research Service forecasts total listed costs for wheat at $419.41 per planted acre in 2026, of which $169.08 is operating cost. That is the lowest total of the nine field crops ERS covers — below barley ($526.92), soybean ($683.89) and maize ($935.79).

Three qualifications travel with that number and are usually stripped off it. It is a forecast for a year that has not finished. It is a United States national average, and ERS publishes regional breakdowns separately. And it is a cost with no revenue beside it, so nothing about margin follows from it.

The variety registers hold a name problem

Official plant variety registers publish wheat varieties by denomination. They also publish strawberry, bean and ryegrass varieties by denomination, and the denominations collide: in United Kingdom plant breeder's rights, the bread wheat name "Cadenza" returns a granted right for a strawberry. Across four registers, 57% of exact name matches to a cultivar belonged to a different botanical species.

There is a second, quieter problem specific to wheat. Registers publish durum at species rank as Triticum durum Desf.; the corpus uses subspecies rank, Triticum turgidum subsp. durum. Both are correct. A matcher that does not know they are the same plant discards genuine registrations, and one that guesses too freely re-opens the homonym door. AgricultureID handles it as a single written-down equivalence.

The trade sources hold two different wheats

Wheat crosses a border twice in the statistics: once as the exporter records it, once as the importer does. The two figures differ, and the difference is not error. They are measured at different borders, at different times, with different valuations, and often under different commodity codes after a classification revision.

Where a reporting country did not report at all, the absence is not a zero — a gap in a trade matrix is a gap, and filling it with zero manufactures a claim that no wheat moved.

The plant health sources hold a list of what to keep out

Wheat appears in quarantine and regulated-pest listings, which are statements about what an authority has decided to regulate in its territory. A pest recommended for regulation in a region is not a pest present in a country, and a listing is not a report of an outbreak. These are three different claims that use overlapping vocabulary.

The soil surveys hold something that is not about wheat at all

A soil survey records what a mapped soil body is like. It says nothing about wheat. The connection between a soil's drainage class and a wheat crop's performance is agronomy, and it is real, and it is not in the survey — a survey that appeared to say "good for wheat" would be a reader's inference wearing an official source's authority.

What this adds up to

There is no source that holds "wheat". There are sources that hold a forecast cost, a set of registered denominations, two asymmetric trade flows, a regulatory list, and a soil description that a person connects to wheat.

The canonical AgricultureID entry is where the crop itself is described. What the Journal can add is this: a sense of which question each source can actually answer, so that a figure quoted from one of them is quoted for the thing it measures.