In Plain Sight
Some growers treat a field as an experiment they are running. They are not gardening by feel. Varieties get compared, yields written down and failures looked at without any sentiment, season after season. You'd notice this in the one whose notebook says which beans came up, and in which row, in 1994.
The Pattern at Work
There are four varieties in the trial and they are not planted in four blocks, which is what everybody does and which settles nothing, because the far corner of that field is wetter. They are planted in strips across the gradient, repeated three times, so that a difference in yield is a difference between varieties rather than between corners. That single decision, made in about ten seconds at sowing, is the whole difference between twenty years of data and twenty years of impressions.
The notebook records what failed and why it might have, in the same detail as what worked. Most growing knowledge is lost at exactly this point. A bad year gets attributed to the weather, which is usually half true and entirely useless, and the row is forgotten. The same mistake gets made eleven years later by the same person, who does not recognise it as the same mistake.
Over about fifteen seasons a picture assembles that no single year could produce — that this soil wants a later sowing than the guidance says, that one variety is ordinary in a good year and much better in a wet one, and that the apparently best performer is only best when the spring is dry, which happens in about half of years and had been quietly flattering it for a decade.
None of that is visible from any single harvest. All of it is worth more than anything a supplier's leaflet says, because the leaflet was written about somebody else's ground.
What the Examples Show
It gets read as being methodical, or as a scientific turn of mind applied to a garden.
The structure is the thing, and it is unusual rather than merely careful. Anybody can keep records; what is rare is arranging the growing so that the records can answer a question — strips rather than blocks, repeats rather than one of each, the same measurement taken the same way every year.
The second element is treating failure as data. That is genuinely hard, because a bad row is a personal disappointment and the sentimental move is to blame the season and move on. This pattern writes it down with the same attention as the success, which is where most of the learning actually lives.
And the timescale is long enough to change what the work is. Fifteen seasons is most of a working life in one place, and the value accrues to whoever inherits the notebook.
Going Deeper
Agricultural improvement has always come from people running trials on their own ground, and the formal science arrived late and mostly systematised what farmers were already doing.
The Rothamsted experiments have run continuously since 1843, were started by a landowner and a chemist on private ground, and are still producing results nobody could have anticipated. The Norfolk four-course rotation spread because individual farmers compared results and then talked to each other about them. Vavilov built an enormous seed collection by treating varieties as a resource to be catalogued, compared and preserved, and people died protecting it. Selective breeding of every crop and animal that people now depend on was carried out over many generations by growers keeping informal records and making choices, long before anybody could explain inheritance or would have known what to call it.
What all of it costs is a timescale that outruns the person.
A trial takes years to say anything, and the answer belongs to whoever is farming the ground afterwards, which in practice means most of this knowledge is lost at a change of tenancy or at a death. The records are also intensely local: what this soil wants is not transferable to a farm three miles away, which makes the knowledge valuable and unpublishable. And there is no money in it. Nobody pays for a notebook, the seed companies fund trials that answer their own questions, and the grower who has done this for thirty years has an asset that cannot be sold and usually goes into a skip.
The Image
Strips, not blocks.
Four varieties down the gradient rather than four squares in the corners. It costs nothing extra to plant and it is the difference between a result and an impression.
Almost every informal comparison anybody makes is blocks: two things tried in different conditions, at different times, and then judged as though the conditions had been the same. The conclusion that follows is usually confident and worth nothing.
Where It Stops
Loving gardening is not this, and neither is keeping a diary. A great many records are kept and answer nothing, because nothing was arranged to be comparable.
It goes wrong as never acting. A mind organised around trials can want one more season of evidence indefinitely, and a farm is not a laboratory: at some point the crop has to be planted at commercial scale on a judgement that is not yet fully supported.
It also fails where the system is not stable. On ground being changed by drainage, new machinery or a shifting climate, fifteen years of comparison can be fifteen years of learning about conditions that no longer exist.
Take the plainer explanation first. Any trained agronomist runs trials because that is the training. The test is whether somebody does it unpaid, on their own ground, with no intention of publishing, and keeps it up for a decade.
Where It Pays
Inside a job. Agronomy, horticulture, seed and variety work, agricultural extension and advisory services, research farms, plant breeding, and the technical end of commercial growing. Also conservation land management and rewilding, where the same discipline decides whether anybody ever learns what actually worked or simply asserts it afterwards.
What is being supplied is local knowledge that no published research can provide. Trials are run at institutions on their soil for their questions, and the gap between that and a particular field is where most yield is lost. One person running honest comparisons on their own ground for fifteen years knows things about that ground which nobody else alive knows, and which no institution has any way of acquiring.
Where it pays badly is anywhere a result is needed this season, anywhere the ground changes hands every few years, and in any tenancy or contract arrangement where the grower has no expectation of being there in ten years, which is a great deal of modern farming. Outside one. Allotments, gardens, smallholdings, a village seed swap. The cost worth naming is that the work produces nothing to show for a decade, and the notebook is worthless to everybody except the next person on that exact ground.
Try This
Pick something you grow, cook or make where you have a choice between two options — two varieties, two methods, two settings.
Do not try one this year and the other next. Do both, at the same time, in conditions as similar as you can manage, and then do it again.
Write down the result and the date, in the same words each time, including what went wrong.
Three repeats is where knowing starts. Most people have never done two.
Strips, not blocks is a tool, not a self. Pick it up where a slow system will answer a question if it is asked properly. Put it down where you need to commit this season, because there the answer is a judgement and no amount of data will arrive in time.
If This Isn't You
Most people plant what did well last year, adjust by feel, and get a perfectly good crop, which is how nearly all growing has always been done. Feeling no urge to keep a notebook is an answer, not a sloppiness.
Where To Go Next
Its near-twin — Ground Yield Maker. Both read soil and season as terms to work within. Ground Yield Maker is building a business and sequencing against year eleven. This one is running a comparison and may have no commercial interest at all.
Its shadow — Field-Naturalist. Field-Naturalist observes a place without altering it and tests explanations against what is already happening. This one changes the field on purpose in order to find something out.
Most often confused with — Plant-Hand. Both work ground by hand across seasons. Plant-Hand is growing to feed people and reads seedlings through the fingers. This one would accept a worse harvest for a cleaner comparison.