Pruning Fruit Trees: Deducing Instead of Memorising

What wood does it fruit on: the question that replaces the calendar

No fruit tree is pruned on a fixed date. It is pruned according to two answers, and those two answers are enough to work out the right cut on any species at all, including a variety you don’t recognise. The first: what wood does the tree carry its fruit on, wood that lasts several years or wood that has to be rebuilt every season? The second: how well does it heal a wound at any given point in its cycle? A calendar built species by species solves both questions for one particular climate and falls apart the moment you cross into another hemisphere; the biology underneath it never changes.

The first split sets the pip fruits, apple and pear, against the stone fruits, cherry, plum and peach. Both groups fruit on short, permanent shoots called spurs, but they don’t heal the same way: pip fruits close a wound made during dormancy cleanly, stone fruits close that same wound poorly and are exposed, right at that moment, to two diseases that get in through the very cut. This one criterion, independent of what wood carries the fruit, is enough on its own to explain why an apple tree is pruned while dormant and a cherry tree while the sap is running.

The second split concerns the fruiting wood itself, and it covers at least three different cases. On apple, pear, cherry and plum, the fruit forms on spurs that stay productive for several years: pruning means renewing part of that stock year after year, never rebuilding it all at once. On the peach, nothing of the sort applies: it fruits only on wood that grew the previous year, a shoot that has already carried peaches will never flower again, and a large share of its one-year wood, not its framework, has to be rebuilt every season, or the harvest retreats toward the tips of the branches. On the fig, the grapevine and the kiwi, a third pattern takes over: the fruit forms on the current season’s growth, but that growth itself starts from a bud carried on one-year wood, which shifts the question entirely. You are no longer choosing between old wood and young wood, you are choosing how many one-year buds to keep.

Getting either question wrong almost never kills the tree, but each mistake leaves its own signature. Getting the healing wrong produces gumming that oozes from the trunk the following season. Getting the fruiting wood wrong on an apple tree left without renewal produces biennial bearing, that pattern of one outsized crop and almost nothing the next. Getting the fruiting wood wrong on a peach produces a harvest that migrates a little further out toward the branch tips every year, with smaller and smaller fruit. The rest of this guide takes these two questions in order, then applies them species by species.

Pip versus stone: why healing dictates two opposite seasons

Apple and pear tolerate being cut in full dormancy for a precise reason: their main health risk at the wound, fire blight (Erwinia amylovora), is itself largely inactive during that period. This bacterium spreads by insects, rain and tools during the growing season, but through dormancy, disinfecting tools isn’t even necessary, the transmission risk is so low. Better still: when a shoot or a canker shows active signs of the disease, sanitation pruning is done precisely during dormancy, specifically to avoid spreading the bacterium over a tree still in leaf. For pip fruits, dormancy is therefore doubly the right window: the wood heals better in the cold than stone-fruit wood does, and the pathogen that threatens them most is asleep at the same time as the tree.

Cherry and plum, like every Prunus, reverse that calculation because of two distinct diseases that both get in through a fresh wound. Silver leaf disease (Chondrostereum purpureum) releases its spores mainly during the damp spells of dormancy, and a cut stays vulnerable to infection for about a week after it’s made. As long as the sap isn’t moving, nothing pushes the spores back out of the wound; once it rises again, during active growth, it tends instead to flush them out. That’s a complete reversal of the apple tree’s case: here it’s active sap that protects the wound, not rest.

Bacterial canker (Pseudomonas syringae) follows a related but distinct logic: the bacterium gets in through pruning wounds and leaf scars during dormancy, then stays latent until the following season, when it shows up as oozing cankers, the amber gumming that gives the disease its common name. A cut made early, before growth resumes, is explicitly listed among the factors that favour its establishment. Gumming itself is not a disease but the tree’s reaction to a wound or an infection: it can appear without bacterial canker, but when it follows a dormant-season cut by a few weeks, bacterial canker is the most likely cause.

Both diseases share the same practical conclusion: prune cherry and plum while in leaf, preferably in dry weather, shortly after harvest or during the height of the growing season, never during winter dormancy. The table below sums up what actually travels by way of the pruning blade, what travels through the wound itself, and what travels by neither.

Brown rot (Monilinia), which browns and rots the fruit before mummifying it on the tree, follows a third logic, independent of pruning timing altogether: the fungus overwinters in those mummified fruits and in shoot cankers, and that’s where it sets off again the following season. Systematically removing mummified fruit and cankered shoots at every visit, in dormancy as much as in leaf, cuts down the stock of spores available for the next season, whatever calendar you otherwise follow for the rest of the pruning.

Reading it for yourself: spurs, spur shoots, one-year wood and the right cut

Before cutting anything at all, a glance at the buds almost always settles the question of fruiting wood. A growth bud is flat, pointed and pressed close against the shoot; a flower bud, on both pip and stone fruits, is bigger, rounder, often noticeably swollen, sometimes carried on a small cushion of wood. On a branch with established spurs, that swelling is visible from several paces away. One-year wood, for its part, is generally recognisable by bark that’s smoother and often redder or paler than wood two years old and older, which is already grey, rough and sometimes fissured: the difference is obvious at a glance when you compare two shoots side by side on the same branch.

On apple and pear, three words name three stages of the same organ. The spur shoot is a very short shoot, often tipped with nothing more than a growth bud, that can turn productive in later years without you needing to create it yourself. The spur is the knotty swelling that forms after a first fruiting, at the base of the old fruit stalk, and it’s this structure that comes back every year, carrying now a flower, now a plain shoot. The spur branch is a fruiting shoot a little longer than a spur shoot, carrying several spurs staggered along its length. None of these words is essential to prune well, but they stop you mistaking a promising spur shoot for a plain growth bud that will never flower, and so removing it by mistake while thinking you’re tidying the tree.

Two families of cuts produce opposite effects, and confusing them accounts for a good many badly shaped trees. A thinning cut removes an entire shoot or branch at its base, at the point where it joins the wood that carries it: it triggers almost no regrowth at that spot, and it opens the tree up to light and air. A heading cut removes only the tip of a shoot, just above a retained bud: it removes that terminal bud’s dominance and triggers a cluster of new shoots just below the cut. On a mature fruit tree, thinning stays the default cut; heading is reserved for three specific uses, thickening a scaffold branch you want to keep, removing something that can’t be left as it is, or deliberately forcing new spurs on a species that fruits on that kind of wood.

The cut itself is made about five millimetres above a bud that faces outward, never toward the centre of the tree. Any closer and the bud dries out; any further and the remaining stub dies and rots before the tree has had time to grow over it. The angled face slopes away from the bud, so rainwater runs off away from it rather than pooling there, a real detail but well below the distance and cleanness of the cut in importance. For a whole branch, cut just in front of the collar at its base, without cutting into it: healing starts from that collar, and removing it along with the branch opens the trunk to rot. Wounds aren’t painted as a general rule, since sealants slow the wood’s natural compartmentalisation rather than helping it; the one documented exception concerns precisely cherry and plum, where a sealant applied right after the cut can delay the onset of silver leaf disease long enough for natural healing to take over.

Apple and pear: the life of a spur, the one-two-three rule and biennial bearing

Apple and pear carry most of their crop on spurs established on wood two years old and older. A spur can physically live for fifteen to twenty years, but it stays genuinely productive for only about eight to ten: past that point it keeps flowering but gives smaller, fewer fruit. A tree that’s never pruned again therefore ends up covered in ageing spurs, with no visible symptom giving warning before the drop in yield itself. The table below sums up, species by species, the fruiting wood, timing and severity for the five major pip and stone fruits in this guide; apple and pear are covered in detail right after it, cherry, plum and peach in the sections that follow.

The simplest safeguard has a name: the one-two-three rule. It means keeping, across the whole tree, a roughly even mix of one-year, two-year and three-year fruiting wood, rather than letting every spur age at the same pace. In practice, that means heading back a small share of the one-year shoots each year to force new spurs at their base, and thinning out the oldest, sparsest spurs. This rule costs less time than managing spur by spur, and it avoids the opposite result, a tree entirely covered in tired old spurs that all crop at once, which worsens biennial bearing rather than correcting it.

Biennial bearing itself comes down to a mechanism of internal competition. Next year’s flower buds differentiate during the summer that follows the current flowering, exactly when the tree is feeding its present crop. A heavy crop exhausts the reserves available for that differentiation: the year after a big harvest therefore mechanically sees fewer flower buds form, giving a light crop, which in turn lets the tree rebuild its reserves and sets up another heavy year. Pruning doesn’t remove this mechanism on its own, thinning young fruit after fruit set works the same lever by a different route, but regularly renewing the spurs limits its swing, and pruning a little more generously right after a heavy-cropping year helps rebalance the tree for the one that follows.

Cherry and plum: the same architecture, a green-season calendar

Cherry and plum carry their fruit on the same kind of spurs as apple and pear, established on wood two years old and older, with cherry spurs staying productive for up to about ten years. The architecture doesn’t change, then: the only thing that changes is the timing, for the healing reasons covered above. Once the framework is formed, these two trees actually need less annual pruning than an apple tree, not more: most of the work is removing dead wood, rubbing branches and upright water sprouts that steal light from the spurs, done during active growth and in dry weather rather than in rain, which is precisely what favours the spread of silver leaf spores and lets bacterial canker in.

On an older cherry tree, spurs eventually crowd each other so closely that they start competing; thinning them out to leave a spacing of around ten centimetres restores size to the fruit that remains, at the cost of some crop volume. The safest timing stays the weeks right after harvest: the wound then has the whole rest of the growing season to heal before dormancy sets in, which keeps it clear of the two diseases covered above. On plum, the same principle of restraint applies, with particular watchfulness over water sprouts, which often push up in numbers at the base of the large branches after a slightly too generous pruning and are best removed early, as soon as they appear, rather than left to turn into wood you’ll have to take out with a saw.

The one exception to this restraint concerns the young tree still being trained, the first three to four years, when the framework is still under construction and more structural cuts remain necessary, always within the same healing windows. Once the framework is in place, restraint becomes the rule again, and the most worthwhile move isn’t even a cut: systematically removing fruit mummified by brown rot and shoots carrying an active canker limits the inoculum available for the following season, far more effectively than harder pruning would.

The peach: the only one that demands a full rebuild every year

The peach breaks with everything above. It fruits only on wood that grew the previous season, never on a permanent spur: a shoot that has carried peaches will never flower again, whatever the age of the tree. Without renewal, the consequence is mechanical and progressive, the harvest retreats a little further toward the tips of the branches every year, to wherever the only wood still young enough to flower happens to be, with smaller and smaller fruit because the shoot carrying them keeps lengthening and thinning instead of staying short and vigorous. Nothing in the tree’s general appearance flags the problem for several seasons: it keeps growing, keeps flowering at the margins, and simply yields a little less each year.

The safeguard is to deliberately rebuild the tree every year rather than suffer that retreat. Each pruning removes on the order of 40 to 50 percent of the wood grown the previous season, a far higher figure than on any other fruit tree in this guide. The precise move is, on each scaffold branch, to pick out a one-year shoot that’s vigorous but not excessive, well placed toward the inside or underside of the branch, keep it as a replacement shoot, then cut off, just above it, the wood that has just fruited. The following season, it’s this replacement shoot, now itself one-year wood, that will carry the crop, and the cycle starts again.

The timing falls a little later in dormancy than for apple, once the flower buds are clearly distinguishable, tinged pink at the tip, but before they fully open. This wait isn’t excessive caution: peach flowers open on still-bare wood and stay exposed to a late cold snap, and the cut itself reduces, for a while, the frost tolerance of the wood that’s just been pruned, which further argues against pruning right before cold weather is forecast. Waiting until the buds are visible lets you spot any damage before deciding which shoots to keep as replacement wood, rather than cutting blind and losing, in the same stroke, healthy wood and the very marker that would have shown you which wood was already affected.

Pruning this hard naturally wakes many dormant buds below every cut, exactly as on any tree cut back hard: most of these shoots need to be removed or pinched out during the season, or the tree closes back up into a needlessly dense mass of wood the following year, instead of concentrating its vigour on the handful of replacement shoots actually chosen.

Fig, grapevine and kiwi: fruiting on the current season’s growth

These three species share a mechanism distinct from everything above. The fruit forms neither on a permanent spur nor directly on one-year wood: it forms on the current season’s growth, itself arising from a bud carried on wood at least a year old. The question is no longer choosing between young wood and old wood, but deciding how many of those buds to keep on the one-year wood, since each one will give a shoot, and each shoot will carry, or not, a flower and then a fruit, depending on the species and its position.

The fig pushes this nuance furthest, with two possible crops on the same tree. A first, early crop forms on wood that has come through dormancy, that is, on still-intact shoots from the previous season. A second, later crop forms on the current season’s growth, and it’s the more reliable of the two, especially where the growing season is short or where winter dormancy damages exposed wood: this second crop forms on new wood regardless, whatever the severity of the pruning. Cutting back hard at the end of dormancy therefore removes the first crop, since no one-year wood is left to carry it, without threatening the second. Light pruning, which preserves some of the previous year’s wood, keeps both crops at the cost of a bulkier tree. The choice depends on what you’re prioritising, not on any universal rule.

The grapevine is trained either in spurs, sections of one-year wood cut back to two or three buds on a permanent framework, or in canes, longer sections carrying more buds on a previous year’s cane left largely intact. Either way, every bud kept will give a shoot that carries the season’s bunches, which reduces the whole decision to a matter of counting rather than a choice between two kinds of wood. One-year wood is generally recognisable by bark that’s smoother and more russet than older wood, which is already grey and rough, letting you sort them even without a label.

Kiwi follows the same logic as the grapevine, with one extra calendar constraint specific to this kind of woody vine: a cut made too late in dormancy, once the sap has already risen, bleeds heavily and can weaken the plant for several weeks. Pruning is therefore done in deep dormancy, well ahead of bud break rather than as it approaches, the opposite of the peach, which prefers to wait as late as possible within its own dormancy window.

Cane and bush fruits: a renewal cycle that varies from one extreme to the other

Summer-fruiting raspberries crop once, on canes grown the previous season. Right after harvest, cut to ground level the canes that have just fruited, recognisable by their duller bark and the dry remains of fruiting, and keep only the most vigorous of the young canes produced that year, at a rate of about six to eight per thirty centimetres of row, removing the rest.

Autumn-fruiting raspberries fruit at the tip of their own cane, the same season it grows, with no need for older wood. The simplest approach is to cut all the canes back to ground level during dormancy, which gives a single crop, late in the season but concentrated and reliable. A double-crop approach exists, shortening only the already-fruited portion of a retained cane to get an earlier second flush the following year, but it complicates management for a gain that isn’t worth it in an ordinary garden.

Redcurrant and blackcurrant look similar enough that gardeners reflexively prune them the same way, which is a mistake in both directions. Redcurrant, like gooseberry, fruits on spurs that establish themselves on a permanent framework of two- to three-year-old wood, almost like a miniature apple tree: prune it lightly, keep its framework year after year, and remove only wood older than four years, which has become unproductive by then. Blackcurrant, by contrast, fruits mainly at the base of one-year wood and on the spurs of two-year wood, and its productivity drops fast beyond that: keep no stem past its third season, in a mix continually renewed from the base, closer to a stool under constant rejuvenation than to a miniature fruit tree. Pruning blackcurrant like redcurrant leaves it clogged with old, unproductive wood; pruning redcurrant like blackcurrant destroys the very framework that carries its fruit.

Blueberry gives its best fruit on wood two to four years old; past about eight years, a cane loses productivity noticeably, signalled by bark that thickens and starts peeling away in strips. Each year, during dormancy, remove about a quarter to a third of the oldest canes, tackling the thickest and greyest first, which keeps up continuous renewal without ever stripping the bush bare in one go.

Elder grown for its fruit gives its best clusters on young wood, its productivity declining sharply after two to three seasons. Each year, during dormancy, remove the oldest stems at the base, aiming for a balanced mix of one-, two- and three-year wood rather than a permanent framework. An elder left without renewal for years is brought back into shape the same way as any neglected fruit tree: the work is spread out, removing its largest stems over several seasons rather than all at once.

Hazel follows a different logic again, because what matters here is neither a spur nor a cane but the catkin, that long, dangling male flower that releases pollen, formed on one-year wood and carried right through dormancy; the tiny female flowers that will become the hazelnuts open in the same place, on that same wood. Removing too much of this one-year wood therefore means wiping out the tree’s pollen supply for the following season in one go: limit yourself to about a quarter of the catkin-bearing wood at each pass, opening up mainly the centre of the bush, most often trained as a multi-stemmed stool, so light reaches the developing nuts. The safest timing falls once the catkins have opened and shed their pollen, rather than while they’re still closed and carrying it.

Rejuvenating an old, neglected fruit tree

The diagnosis needs no tools: a fruit tree left several years without pruning carries its crop higher and further out every year, around the edge of the crown, while the centre, starved of light, carries nothing but dead or unproductive wood. Some scaffold branches may have started competing with the trunk, others cross and wound each other in the slightest wind. None of this kills the tree, but all of it reduces its crop year after year, and the instinct to cut it all back hard in one go to make up for lost time is exactly the move to avoid.

A single brutal, sweeping cut doesn’t restart fruiting, it triggers the opposite: hard heading back wakes a mass of dormant buds all at once, and the tree responds with a hedge of vigorous upright shoots, water sprouts, clustered just below the big cuts. These shoots are typically barren the first season, sometimes the next one too, and the tree can spend several years rebuilding wood before it fruits normally again. Cutting it all back at once therefore trades a mediocre but real crop for several seasons with no crop at all.

Rejuvenating the tree instead spreads over three dormant seasons. In the first, remove dead wood, diseased wood and the worst faults in the framework, crossing branches, competitors with the trunk, without exceeding about a third of the volume that needs correcting. In the second, once you’ve seen how the tree has reacted, thin out the water sprouts that have appeared in the meantime, keeping only a few well-placed ones to fill a gap or replace a failing scaffold branch, and continue opening up the centre. The third finishes the framework. This staged approach applies to apple, pear, cherry and plum, whose old wood remains capable of fruiting again once it’s opened up to light.

The peach escapes this caution, for a reason already covered: its old wood will never bear fruit again regardless, renewed or not. On a neglected peach tree, the question isn’t how to stage the rejuvenation but accepting that a good part of the framework has to be rebuilt in one go to force out new wood, or the tree stays unproductive indefinitely, however much patience you bring to it.

A tree that has just been through hard rejuvenation pruning carries a root system out of proportion to its now-reduced crown: it benefits from regular watering through the first season that follows, while it rebuilds its foliage. Nitrogen applied at the time of cutting only worsens the reaction, feeding the surge of wood growth even more at the expense of the return to fruiting that the whole operation is meant to achieve.

Common pitfalls to avoid

The care sheets to read next

Every plant mentioned has its full sheet: watering, light, soil, pests and a month-by-month calendar.

Browse every sheet in the catalogue →