AS a plant breeder by training, I know the seductive power of a beautiful number. I also left plant breeding with one lesson firmly planted in me: breeding demands extraordinary patience.
You make a cross, select what looks promising, plant it, wait, measure and wait again. By the time a truly improved palm proves itself, the breeder who began the journey may already have moved on.
Perhaps that was partly why I eventually wandered into plantation management, corporate and advocacy. I was probably too impatient to wait for the grandchildren of my crosses to tell me whether I had been clever.
The oil palm works on biological time and has little respect for quarterly reports. Yet the breeder in me never quite disappeared.
That is why I found work from KLK Sawit Nusantara’s Quality, Training & Research Centre (QTRC), presented at the Mosta Oil Palm Best Practices Workshop in Sandakan, so thought-provoking.
The presentation was led by Siti Mashilah and Sharifah Asmahani with their QTRC colleagues.
Siti now manages QTRC, supported by Sharifah, while plant breeder and agronomist Christopher Donough serves as adviser. The breeding programme traces its roots back about 25 years, to when the operation was still part of IJM Plantations.
Behind all the measurements sits a deceptively simple question: should breeders concentrate only on producing palms capable of higher yield, or should they also breed palms that make it easier for growers to recover more of that yield?
That distinction matters because growers do not harvest genetic potential. They harvest bunches.
Between what a palm can biologically produce and what finally appears in an estate production report lie soil, rainfall, palm age, nutrition, labour, harvesting rounds, loose-fruit recovery, crop evacuation, supervision and milling. Potential has to survive reality.
The difference between promise and payment
Oil palm scientists have long distinguished between theoretical yield, experimental yield and what is actually achieved commercially.
Detailed research-plot records suggest that peak fresh fruit bunch (FFB) potential could be around 50 tonnes per hectare or more under favourable circumstances.
Nobody, however, is suggesting that an estate should simply buy better seed and expect 50 tonnes to arrive obediently at the weighbridge.
The reason is familiar: realised yield depends on genetics, environment and management. Excellent planting material on difficult soil will not perform as it might elsewhere, while good genetics without good agronomy will disappoint.
A palm may manufacture a magnificent ripe bunch, but if nobody cuts it in time, part of that biological achievement never becomes commercial yield. Potential, in other words, is not a cheque that automatically clears.
Malaysia’s national figures make that distinction relevant. Using the Malaysian Palm Oil Board statistics through 2024, the presentation notes average FFB yield of around 19 tonnes per hectare over 1990 to 2014, about 17.2 tonnes during 2015 to 2019 and around 16 tonnes over 2020 to 2024.
Explanations include ageing palms, delayed replanting, inadequate harvesting labour and difficult weather.
Those explanations are familiar. What is less often asked is whether breeders can help make some operational difficulties less severe.
When harvesting falters, we call it a labour problem. When ripe bunches remain uncut, we call it management. When loose fruit lies on the ground, we call it field loss.
All true. But perhaps some management problems can also become breeding opportunities.
Yield-making and yield-taking
One of the QTRC team’s most useful distinctions is between yield-making and yield-taking.
Yield-making happens inside the palm. Leaves capture sunlight and convert it through photosynthesis into carbohydrates, the raw material for roots, trunk, fronds, flowers, fruits and ultimately oil. Physiological work describes oil palm as largely “source-limited”: there is only so much photosynthetic capacity available for these competing demands.
Think of the palm as a household living on a salary. The same ringgit cannot be spent twice.
Oil is particularly expensive, containing roughly twice the energy of the other major bunch components, so breeders cannot assume bunch yield and oil concentration can simply be pushed upward indefinitely together.
Mother Nature, it seems, also has a finance department.
The industry naturally wants everything: more bunches, higher oil content, fewer losses, easier harvesting, shorter palms, better stress tolerance and lower costs.
The breeding wish list can begin to resemble an all-you-can-eat biological buffet. But plants are not buffet counters.
The QTRC team’s practical emphasis is therefore on recoverable and recovered FFB.
Without the bunch reaching the mill, there is no additional oil to extract. That takes us from yield-making to yield-taking, and suddenly the breeder has to think like the person holding the harvesting pole.
Breed for the harvester too
Imagine two palms. Palm A has spectacular theoretical yield but carries heavy fronds that are difficult to cut, sheds ripe fruit rapidly and produces a huge seasonal crop peak precisely when the estate is already struggling for harvesters.
Palm B may look slightly less impressive on the breeder’s spreadsheet but has lighter fronds, accessible bunches, fruit that is easier to recover and a more even crop distribution through the year.
Which palm is better? That depends on whether you are judging a research plot or running an estate.
Among the traits examined by the QTRC team are smaller petioles producing slender, lighter fronds, making cutting easier while retaining useful leaf area.
Larger fruitlets may simplify loose-fruit collection, while better distribution of bunch production through the year could reduce extreme peaks and troughs and help harvesters maintain productivity.
They are also looking at slower ripening, slower quality deterioration after harvest, and palms combining high oil content with slower fruit detachment.
The approach is refreshingly realistic: improvement is likely to be incremental rather than a quantum leap.
To somebody outside plantations, these may sound like small botanical details. To the person doing the work, they are not small at all.
A frond is not merely plant architecture; somebody has to cut it. A loose fruit is not merely evidence that a bunch is ripe; somebody has to collect it. A crop peak is not merely a curve on PowerPoint; it means the same workforce suddenly has much more crop to recover before losses rise.
And a ripe bunch left hanging on the palm contributes precisely zero tonnes to the estate production statement. This is why the breeding objective begins to shift from merely producing the highest-yielding palm towards producing a palm whose yield can actually be exploited.
There is another lesson hidden inside the fruit itself. Oil palm produces not one commercial oil but two: crude palm oil from the fleshy mesocarp and palm kernel oil from the kernel inside the nut.
Historically, breeders have measured crude palm oil-related characteristics in great detail, while the actual oil content of individual kernels has received much less direct attention.
Kernel-oil yield has commonly been estimated from whole kernel yield using a standard conversion rather than measured directly in prospective breeding parents.
That suggests another breeding principle: what we do not measure, notice or value is unlikely to find its way onto the breeder’s wish list.
A preliminary QTRC study added kernel-oil extraction to normal bunch analysis and found meaningful palm-to-palm variation in kernel oil content among the DxP palms studied, with variation also observed among Deli dura parents.
Because this had not previously been a selection target, wider measurement may eventually help identify parents combining desirable crude palm oil and kernel-oil characteristics.
That does not mean a high-palm kernel oil “super palm” has suddenly been discovered. The work is preliminary, and variation alone does not tell us how reliably a trait will be inherited or what trade-offs may accompany it.
The larger lesson is simpler: progress does not always begin with discovering something entirely new; sometimes it begins by looking again at something familiar through a different lens.
Breeding advances not only by finding better answers, but by learning to ask better questions.
The palm does not read the research protocol
Another important part of the QTRC approach is that breeders must work hand-in-hand with growers.
Research plots are indispensable. Good breeding requires controlled comparisons, accurate measurements and painstaking records accumulated over many years.
But no research plot can reproduce everything that happens across thousands of hectares of commercial plantation.
Commercial agriculture is untidy. Rain falls when the spreadsheet did not request it. Workers leave. Roads wash out. Harvesting rounds stretch. Soils change. A mill occasionally chooses precisely the wrong day to have a problem.
The plantation does not read the research protocol.
That means breeding material ultimately has to be observed in commercial plantings, under the conditions and scale at which growers actually operate.
Performance must be benchmarked against what happens in the real world, where data may be messy but the lessons can be invaluable.
Instead of breeding ending with, “Here is our improved seed; good luck!”, breeder and grower should continue learning together.
The grower discovers how best to exploit the genetics; the breeder discovers what reality does to the genetics.
Agricultural science occasionally suffers from a temptation common to many professions: when reality disagrees with our model, we become irritated with reality. Reality usually wins.
A different report card
Oil palm breeding has traditionally measured bunch number, bunch weight, oil content, mesocarp characteristics and total yield. They remain essential.
But perhaps tomorrow’s breeder needs another report card alongside them.
How difficult is the palm to harvest? How heavy are its fronds? How accessible are its bunches? How rapidly does ripe fruit detach? How evenly is crop distributed through the year? How forgiving is the fruit when harvesting intervals stretch?
Are there valuable characteristics, such as kernel oil content, that we have traditionally estimated rather than directly measured?
Ultimately, these questions return to one larger question: how much of what the palm produces biologically eventually becomes something the grower can sell?
That moves breeding beyond creating a champion tree. It asks whether breeding can help create a champion production system. Better genetics alone will never be enough, but genetics can increasingly be designed to work with better agronomy, harvesting and milling rather than sitting apart from them.
Breeding still wears muddy boots
There is a refreshing absence of miracle language in the QTRC team’s work. Progress is expected to be incremental; breeding remains long, patient and decidedly unglamorous.
The presentation closes with researchers getting “down and dirty in the fields” if growers are to move closer to the biological ceiling.
I rather like that image.
For all the sophistication of modern agriculture, progress may still begin with someone standing in the mud, comparing two palms and asking why one frond is lighter, one bunch easier to cut, one fruit larger or one crop better distributed.
Innovation does not always arrive wearing a laboratory coat. Sometimes it wears muddy boots.
Perhaps that is why, decades after leaving plant breeding, I still admire those patient enough to stay with it.
Breeders plant today for answers that may come years later. Many selections will disappoint, a few will surprise, and occasionally one palm will justify all that crossing, measuring and waiting.
Breeding programmes outlive individual careers. Knowledge is handed on to another generation, which asks new questions of old material and takes the work forward.
The best palm of the future may therefore not be the one with the most spectacular theoretical number, but the one whose potential can be more fully realised in the field.
That does not lower the ambition of breeding; it makes the ambition useful.
A plantation is not paid for what a palm could have produced, but for what is harvested, recovered, milled and sold.
After all those years of crossing, selecting and waiting, perhaps the real breakthrough is not simply breeding a palm with an ever-higher ceiling.
It is breeding one that helps more growers reach it.
Joseph Tek Choon Yee has over 30 years of experience in the plantation industry, with a strong background in oil palm research and development, C-suite leadership and industry advocacy. The views expressed here are the writer’s own.
