THE AMORPHOPHALLUS INFLORESCENCE

Most of what a grower needs is timing. The female flowers of nearly every species in this genus are receptive on the first day, often for a matter of hours, and the same inflorescence sheds its own pollen a day or two later — by which point its own stigmas have shut. That single fact governs everything downstream: it is why you cannot self a plant casually, why a second plant almost never obliges by flowering the same week, and why the practical heart of this article is not the cross at all but Part VIII, on keeping pollen in a freezer until the next inflorescence arrives.

The rest is worth knowing because it tells you what you are looking at. The heat and the stench are not decoration around the reproductive parts; for most of this genus they are the recruitment strategy, and the plant has no second one. When a cross fails, the failure is usually in that machinery rather than in the pollen.

A note on what this page is willing to say

The Amorphophallus literature is unusually honest about its own thinness, and this article inherits that. Where two papers disagree — and on the smell of the titan arum, four of them do — the disagreement is printed rather than resolved by picking a favorite. Where a widely repeated number turns out to be a calculation rather than a measurement, it is labeled as one.

If you grow these plants and keep records of what flowered, when, what you crossed and what came of it, you hold data the published literature does not have. Part XII and the sources block both say where the holes are.

What looks like one enormous flower is a spadix — a fleshy column carrying hundreds of tiny naked flowers — wrapped in a single modified leaf called the spathe. The flowers themselves have no petals and no sepals. A female flower is an ovary with a style and a stigma on top; a male flower is little more than an anther on a stalk. They are unisexual, and they are sorted into bands.

Bottom to top, on the spadix

1 · The pistillate zone. The female flowers, at the very bottom, hidden inside the closed base of the spathe. Each is an ovary with 1–4 locules, topped by a style that may be almost absent or several millimeters long, and a stigma that may be lobed, capitate, discoid, conic or bilabiate depending on the species. This is the zone you are trying to reach.

2 · The staminate zone. The male flowers, in a band immediately above the females. Each carries one to six stamens (rarely eight), free or sometimes fused, and the anthers open by pores or by terminal slits. This is where pollen appears — a day or two after the females below it have already closed for business.

3 · The sterile zone, if there is one. Between the two fertile bands some species carry a ring of staminodes (also called synandrodes or neuter flowers) — sterile structures that are shield-like, globose, hairlike, conic, warty or spiny, according to species. Many species have none at all and the two fertile zones simply touch. Where the ring exists it is not filler: in Amorphophallus konkanensis and Amorphophallus sylvaticus the swollen neuter flowers are the food that keeps beetles inside the chamber for days.

4 · The appendix. The sterile column on top, carrying no flowers at all, and in most species the largest single thing in the inflorescence. This is the scent organ and, where the plant heats, the principal furnace. Its shape is one of the genus's main taxonomic characters — conic, fusiform, cylindric, mouse-tailed, pendulous, deeply corrugated like a brain, or covered in filaments or staminodes.

And around all of it, the spathe

The spathe divides into a base and a limb. The base wraps on itself — the term is convolute — and encloses the fertile zones in a chamber that stays shut while the limb above it opens out. In many species the two are separated by a constriction, a waist where the spathe pinches in; in others the constriction is shallow, and in a good many there is none.

The inside surface of that base is where the plant does its physical work on an insect. It may be smooth, ridged, warty, or lined with hair-like filiform processes — and in several species it is coated with a secretion that makes it slippery enough to fall down and hard to climb out of.

The character states, as the keys record them

The table is here because the genus does not have a standard inflorescence. If you are trying to describe your own plant, or to work out which species you have, these are the axes the literature actually measures it on.

Character states after Scholten 2023 (239 spp.) and the Hetterscheid & Ittenbach descriptor set as enumerated by Anil et al. 2023 Measured

One more thing to notice about that list

Every row is a range, and several of them are ranges that a single species covers on its own. That is not a failure of the keys; it is the genus. Part II is about how much a species varies from itself, and it is the most important caveat on this page.

The scale first. Across the genus, leaves run from 2 cm long in Amorphophallus pusillus to 5 m long and 7 m across in Amorphophallus titanum. Inflorescences run from under ten centimeters tall, in Amorphophallus aphyllus and Amorphophallus terrestris, to a spadix over 100 cm in Amorphophallus titanum on a plant that has been recorded at close to three meters overall. Some species flower almost entirely underground.

Set beside that, the thermogenesis data look luxurious: temperature has been logged in 80 species. That is the best-sampled trait in the genus, and what it found was that a quarter of them do not heat. Part IV takes that apart.

Genus figures after Scholten 2023 · intraspecific data from Anil et al. 2023 · pollen from Punekar & Kumaran 2010 and Raman et al. 2017 · visitor tally from Claudel 2021 · the diploid Amorphophallus bulbifer count from Das et al. 2020

Amorphophallus is protogynous: female first, male second, on one inflorescence. Anthesis usually lasts two days. On day one the stigmas are receptive and the plant does its entire advertising campaign — heat, smell, color. On day two the anthers open and pollen appears, by which time the stigmas below have already gone over.

This is the single fact that shapes the whole practice. You cannot wait for the pollen to appear and then use it, because by then there is nothing receptive left to put it on. Either a second inflorescence is open at the right moment — which almost never happens, because these plants flower once every few years and rarely in step — or you are working from stored pollen. Part VIII is the answer to that problem, and it is why it is the longest method section on this page. A matched pair does flower in step now and then, though. A grower signing as Dewey reported that two Amorphophallus abyssinicus bloomed at the right time and in the proper time sequence in summer 2004, that hand pollination of the second inflorescence set seed and bore fruit, and that by January 31st, 2005 the seedlings were about six inches tall or more, most starting a second juvenile leaf. The post does not say where the pollen came from.⁶⁶

What is known, species by species

Intervals for Amorphophallus johnsonii, Amorphophallus eichleri and Anchomanes are derived here from the published per-inflorescence dates — the pollination date and the date that same inflorescence's pollen was collected. The authors never printed an interval. The arithmetic is theirs; the conclusion is drawn here.

How long the whole thing lasts, which is a different question

The receptive window is hours. The inflorescence lasts a good deal longer, and the two get confused constantly. In Amorphophallus paeoniifolius, measured: the shine goes off the spathe and appendix one day after full bloom; the inflorescence withers at three to five days; it collapses at eight to eleven; and the peduncle detaches from the corm at thirteen, at which point the plant returns to dormancy.

For Amorphophallus titanum the garden account above carries the tail as well: the inflorescence closes completely on the second day, may stand upright about a day longer, and then the spadix collapses — a photograph of that collapse is captioned “4 days after opening”. An inflorescence that was not pollinated “will wither in the following weeks”. No sample size is attached to any of it.

Across the genus, the full thermogenic sequence takes about 48 hours in 74 of the 80 species measured. Four run to five or six days. One — Amorphophallus schmidtiae — has been recorded running for up to three weeks.

Thermogenesis in Amorphophallus is the best-sampled thing about the genus, because one project spent seven years pushing thermocouples into 119 inflorescences across 80 species — against the nine species that had ever been measured before it. What that survey found was not a genus-wide furnace. It was a spectrum with a large cold end.

80 species · thermocouples 2–3 mm deep in the female zone, male zone and lower appendix, logged every five minutes from the onset of anthesis Measured

Bigger is not hotter

The intuitive rule fails outright. Some of the hottest species in the genus are among the smallest; Amorphophallus gigas, one of the largest, shows no temperature rise at all. What the statistics do find is subtler and points the other way from size: hotter species tend to be shorter and thicker — the strongest correlation is with the radius of the male zone, not with any length.

The authors' conclusion is worth carrying: there is no emblematic thermogenic inflorescence, and floral traits cannot substitute for actually measuring the temperature.

Which parts heat, and in what order

There is no canonical sequence either. The textbook pattern — appendix on day one, male zone on day two — holds for a majority, but the survey documented species doing nearly everything else:

An eighty-first species, measured after the survey closed

Amorphophallus koratensis was not among the eighty. It was logged separately in 2024, on a plant grown at home with four thermocouples, and it is worth having because it puts a number on the gap between the appendix heat peak and the male-zone maximum — which almost nothing else in this genus does.

Three close relatives, three different curves. That is the eighty-species survey's lesson arrived at independently: the thermal pattern is a species character, and knowing a plant's nearest cousin does not tell you what it will do.

The same genotype, two different patterns

Clonal ramets usually repeat their parent's thermal signature, which makes one exception genuinely strange. Clones of Amorphophallus schmidtiae — identical genotype, identical morphology — produced two distinct thermogenic types: one running for weeks, one finishing in two days. Scent was noticeable only on day one in the long type. The authors treat this as proof that genotype and shape cannot explain the pattern by themselves, and leave it open.

Degrees, and then watts

Only three species have ever been followed with anything more than a thermocouple, and the numbers do not translate neatly.

A live argument worth knowing about

The respirometry study reported water running down the appendix at peak heat and concluded that evaporation was carrying away most of the output — which would mean temperature readings badly understate what the plant is doing. The 80-species survey disputes it on three grounds: the droplets in other species have been identified as odoriferous secretion, not water; the appendix has a thick cuticle and, in Amorphophallus titanum, no stomata at all; and the species that genuinely do cool below ambient are the ones emitting large amounts of highly volatile trimethylamine, which points at scent release rather than transpiration as the coolant.

It goes further: a high respiration rate at anthesis may not mean heat at all. It may be the cost of making and releasing the smell, which is expensive. The precedent is cycad cones, where exactly that turned out to be the case.

A third observation, from the Amorphophallus koratensis work above, lands on the survey's side. A thermal-camera frame, taken when a thermocouple read 30.8 °C — which the author calls almost identical to the camera reading — put the inner wall of the spathe at 23.8 °C, the coldest thing in the picture, cooler than the room. On that same wall, on the day of peak heat production, the surface was visibly moist and carrying droplets, and a strong sulfurous smell was coming off the plant. The author's reading is the direct one: the liquid was evaporating, and evaporating cooled the wall.

That is a different claim from the original one, and a more careful one. It is not that the furnace is losing its heat to evaporation. It is that a separate, cool surface is shedding a scented liquid at the same moment. The nearest thing to a controlled version of this is work on Amorphophallus gigas, published under the title the cooling effect of the odorous liquid — which is where the survey's argument came from in the first place.

So what is the heat for?

Two hypotheses have been in the literature for decades. The first is that heat volatilizes the scent — and it has real support in the chemistry, because the hottest species in the genus emit benzenoids with very low volatility that would struggle to get airborne cold. The second is that the heat is itself the reward: warming the visiting insect directly, or providing a heated chamber worth sheltering, feeding or mating in.

A further proposal is newer, and narrower than either. Working on the sulfur chemistry of Amorphophallus titanum, the authors of that species' first transcriptome suggest the appendix's temperature speeds the last step of the odorant itself — methanethiol oxidizing spontaneously into dimethyl disulfide and dimethyl trisulfide. On that reading the heat is part of the synthesis rather than only part of the delivery. They put it in a single sentence and measure no kinetics.

The survey's own position is that the function is probably not fixed or universal — that it varies with each species' chemistry and climate, and that the links are “loose and non-deterministic”. Given that a quarter of the genus manages to reproduce without heating at all, that is hard to argue with.

The objection is ninety years old

The scent-volatilizing explanation is usually presented as the settled one, and it is worth knowing that the man who proposed it withdrew it himself. Working in Java in the 1930s, he had argued in 1933 that the heat served to evaporate the smell. Four years later he published the reason it could not be the whole answer, and it was his own data: in one species the male flowers heat and produce no odor at all, and in others the warmth and the smell simply do not coincide in time.

His three species make the point better than any general argument, because they disagree with each other completely.

Three species · Java, cultivated and wild · thermometer bored into the appendix pith for Amorphophallus oncophyllus; hand only for Amorphophallus variabilis; Amorphophallus titanum reported by a correspondent in Sumatra · van der Pijl 1937

Where it came from

Mapped onto the phylogeny, thermogenesis appears to have arisen once, at the base of the genus, and to have been lost independently several times in unrelated branches. It did not drive the genus's diversification — the model in which speciation is independent of thermogenesis fits better than the one in which it is not.

And one measurement from the forest

Everything above was measured in a glasshouse. There is exactly one published set of readings taken from a wild Amorphophallus titanum standing in Sumatran forest, and it is worth knowing precisely because it looks like it contradicts the rest.

So the honest position is that nobody has measured a wild titan arum at its hottest. The greenhouse figures are the only amplitude data anyone has, and they come from plants in pots.

Two details from the thermal images

The two-night scheme has a shape that is easy to miss. On the first night the appendix peaks around midnight. On the second the male florets peak between about 18:40 and 20:00 — four hours earlier in the evening, not later. If you are trying to be present for pollen release, that is the practical consequence.

Around ninety species have had their inflorescence odor analyzed by gas chromatography, and the results sort into seven chemical families. The largest by far is the sulfur group: some proportion of dimethyl oligosulfides turns up in 58 of 92 species, and dimethyl trisulfide specifically in 47. Those are the compounds of decaying protein, and they are what a carrion insect is built to find.

But the remaining categories are not marginal. Whole clades have gone somewhere else entirely, and in several cases a species' scent is one compound at over ninety per cent.

Chemistry from a GC-MS survey of 92 species mapped onto a phylogeny, with earlier headspace work Measured   Most species were sampled once.

Two growers’ noses can be set beside the table. In 1999 the first inflorescence of an accession grown as Amorphophallus haematospadix, listed above with the banana esters, had a fragrance much like rotting fruit rather than the carrion smell of many other species, according to Craig M. Allen, then conservatory manager at Fairchild Tropical Garden.⁶⁸ An Amorphophallus henryi, listed with the dung group, bloomed in a campus greenhouse in central California on May 25th, 2007; its grower, C. J. Addington, described it as short, dark purple and brown, and smelling of dog droppings.⁶⁹ Each is one plant and one nose.

Where the smell sits on the family tree

A separate survey ran headspace analysis on 80 species and laid the results against published molecular phylogenies. Two things came out of it that a grower can use.

The first is geographic. Odors built mainly of dimethyl oligosulfides — the ones described as “gaseous” rather than as rotting meat — were found only among the Asian species, and several of those cluster in particular branches of the tree, including two within subgenus Metandrium. So the smell is not scattered at random across the genus.

The second is that it is not neatly inherited either. Some gaseous-smelling species turn out to be close relatives of rotting-meat species, whose dominant oligosulfides carry a set of minor components alongside them. The authors describe the pattern as both phylogenetic constraint and plasticity, and this page follows them: knowing a plant’s section narrows the guess and does not settle it.

One correlation is worth carrying into the greenhouse. Odor has co-evolved with color: the species that smell of rotting meat have darker inflorescences, while the gaseous-smelling ones are generally paler. Species with pleasant benzenoid odors form two groups that are not close relatives at all. It is a tendency, not a key — but a dark spathe is a reasonable prior for what is about to happen to the air.

And a species can disagree with itself

Repeat analyses of the same species land in different chemical categories. Amorphophallus eichleri has been recorded at 62%, 56% and then 23% dimethyl disulfide, with 2-heptanone climbing 7% → 25% → 30% — the third sample would be classified as an alcohols-and-ketones species rather than a sulfur one. Amorphophallus symonianus came out 60% acetophenone in one plant and 89% 1-phenylethyl acetate in another.

Amorphophallus bulbifer shows the same disagreement in the greenhouse. One grower reported that every Amorphophallus bulbifer in the collection smells like propane gas when it first opens, enough that the smell of an open flower has often been mistaken for a gas leak.⁴¹ Another described the largest Amorphophallus bulbifer in a school greenhouse as giving a touch of natural gas leak with the nose in the flower but, from a couple of feet away, a floral, slightly sweet, subtle scent.⁴² Wilbert Hetterscheid asked whether the pleasant-smelling plants were true Amorphophallus bulbifer or the bulbifer-like Amorphophallus muelleri, whose smell carries additional chemicals.⁴³ He also wrote that he believes clones with a deviating smell may exist, and suggested that the balance of the dominant dimethyl trisulfide against small amounts of two other sulfides could differ between clones; he co-authored an early chemical study of Amorphophallus scents.⁴⁴ These are grower reports, not measurements.

The question whether a plant is Amorphophallus bulbifer or Amorphophallus muelleri has a partial answer in the tuber. Wilbert Hetterscheid wrote that vegetative plants of Amorphophallus bulbifer are difficult to separate from Amorphophallus muelleri, but that the tuber may help: that of Amorphophallus bulbifer is smooth, while that of Amorphophallus muelleri carries distinct, raised, annulate root scars all over its surface.⁷⁰,⁷¹

The titan arum is the extreme case, and it deserves its own box.

A useful detail from the one study that used a human nose

Only one analysis put assessors at the sniffing port alongside the detector, and it found something the chemistry alone would have hidden: dimethyl disulfide, present in quantity, produced no perceptible smell at all. Its threshold is too high. What people actually smell is dimethyl trisulfide — and the disulfide is apparently there chiefly as the trisulfide's precursor.

The same study logged the smell's shift through the night: rotten fruit at first, then something like pickled radish, then rotten egg; a constant rotting-animal smell once the spathe was fully open; and after the temperature peak, once the spadix began exuding fluid, a switch to rotten fish. That last phase is the trimethylamine, and it comes out of the liquid rather than the tissue.

---

The oldest dissent, and it is worth reading

The strongest objection to the carrion generalization is also the earliest. In Java in the 1930s, a botanist working on Amorphophallus variabilis noted that the local name for it, kembang bangke, means carrion flower — and then disagreed with it, in print, standing next to the plant:

One nose, one species, and no instrument — so it settles nothing on its own. But it is a first-hand observation from someone who had the plant in front of him, it agrees with what the fermenting-fruit species in this genus are now known to emit, and it makes the point that the carrion reputation was being over-applied ninety years ago by people who had not smelled the plant they were describing.

Amorphophallus variabilis · Bandoeng, Java · one observer, unaided nose Single observation

---

Where the smell is actually made

The scent-producing tissue is called an osmophore. Only one species in the genus has ever had one described anatomically, and it is worth knowing what that study found because it contradicts the usual shorthand.

The received wisdom is that the appendix is the scent organ. In this species it is the strongest, but not the only one. Secretory tissue was found on the appendix (strongest), the staminate flowers (weaker), and the inner face of the spathe (weak, and only in its upper and middle parts). Vital staining came back positive in all five regions tested — including the spathe base, where a human nose detects nothing.

Structurally the gland is simple: a thin-walled epidermis doing the emitting, over two to seven layers of tightly packed parenchyma stuffed with starch grains, with small vascular bundles running right up against it. Deeper in sit larger, emptier cells that look like a storage tissue, and the spadix's core is a well-developed system of large air spaces — the ventilation a tissue respiring that hard needs.

Under the electron microscope the appendix surface is covered in wax sticks, with two-micrometer droplets pushing through the cell wall in the gaps between them and around the stomata. Stomata are present but sparse and sunk under heavy cuticle — about 8 per mm² on the appendix against 62 per mm² on the outside of the spathe base. Many osmophores in other plant families have none at all, so their role here is probably contributory rather than essential.

Made in one place, released in another

For Amorphophallus titanum there is an argument that those are two questions with two different answers. A molecular study of that species — cultivated plants, sampled tissue by tissue through the night — puts the manufacture of the disulfides in the appendix, which is where the sulfur pathway and the enzyme that makes methanethiol in a single step are both running on the day. But it reads the appendix's thick outer cuticle as throttling emission there, and puts the release somewhere else: the spadix is hollow from the appendix down through the ring of male flowers, and the authors have the volatiles circulating inside that ring and leaving from the basal cup of the spathe rather than off the outer surface of the appendix. That points at the same exit as the field description of the plume below, and neither account cites the other.

The plume is not a sphere

Everything above concerns what the smell is made of. There is one published account of where it goes, and it comes from the same Sumatran fieldwork that produced the only wild thermometry in Part IV.

The scent did not surround the plant. It came in waves and only in certain directions, always issuing from the overlap of the spathe — and almost always downhill, where it was “perceptible even several meters away”. Uphill it was “nearly undetectable.” In both plants observed, the spathe overlap faced downhill.

Which insect is each smell aimed at?

Almost entirely a matter of inference — but not quite. One species has been taken as far as the antenna. In Amorphophallus albus, a gas chromatograph run alongside an electroantennographic detector showed that the antennae of its rove-beetle pollinators — Atheta sp. and Anotylus sp. — respond clearly to dimethyl disulfide and dimethyl trisulfide, the two compounds that make up over 92% of that species' blend. That is the only measured link in this genus from a named compound to a named insect. Every other match on this page is drawn by analogy — from Arum, from blowfly baits, from the chemistry of fish decay, from what is known to attract beetles in other systems.

The analogies are reasonable ones. Dimethyl disulfide is an ingredient of commercial screwworm bait. Trimethylamine is the marker of decaying fish and a known housefly attractant. 4-methoxyphenethyl alcohol — the anise compound in the hottest species — is a strong beetle attractant elsewhere. But 1-phenylethyl acetate, the whole scent of the sweet-smelling clade, is not a known attractant for anything, and nobody knows what visits those species.

What a grower can take from it: the smell is a species character worth recording. If your plant smells of something other than death, that is information — and given how thin this literature is, it may be information nobody has written down.

Start with the shape of what is known. Insect visitors have been reported for around two dozen Amorphophallus species, under a tenth of the genus. About a third of those reports come from a single inflorescence in the wild — one plant, one night, one observer. And the outcome that actually matters, whether fruit was set, has been quantified in the wild for a handful of species.

Against that, the family-wide picture is that Amorphophallus is comparatively well studied. It is one of a handful of aroid genera with all three major insect orders on record, and it carries eight beetle families — the highest count in the family. Fifty-six aroid genera have no pollinator records at all.

The three grades, and why they are kept apart

Throughout this section a chip says how good the record is. Pollinator means the insect was shown to carry pollen from one inflorescence to a receptive zone on another. Putative means it was seen in the reproductive zone, doing something plausible. Visitor means it was on the plant. The published lists mostly do not sort these, and the compilation that lists them all says so in its own text: only a few — sometimes only one — are the real pollinators of any given species.

Sorted after a 2021 review that split every published record into pollinator, putative and visitor. Where that review and the primary field papers disagree, this table follows the primary.

The trap, and whether it is really a trap

Several species hold their visitors inside overnight, and the machinery has been described in detail for a few. A field study in western India laid it out as five successive traps: a visual one — the shape and color of the spathe and the height of the peduncle; an odor one; a slippery one, where the wet appendix and the inner spathe wall send the insect sliding to the bottom; a food one — stigmatic fluid, sterile flowers, appendix tissue; and finally a reproductive one, a dark sheltered place to mate and lay eggs in.

The mechanism of the slippery phase, where it has been looked at, is hairs: the inner spathe base carries rounded warts bearing unicellular hairs that “secrete certain chemicals that make the basal inner part of the spathe slippery”. An overhanging edge at the base of the appendix, sharp enough that the beetles fall back down every time they try it, has been described for one species only, Amorphophallus titanum — and that account is second-hand, relayed in 1937 from a collector who never watched the pollinators arrive and who called the trap no more than probable.

An idea worth having in your head while you watch

One review floats a hypothesis that reframes the whole guest list: that some of these insects are not fooled at all, and are not there for the imaginary carcass. They are there to hunt the insects that were.

The evidence is circumstantial but pointed. Creophilus rove beetles are named as the exclusive pollinators of one species — and Creophilus are predators, feeding on carrion-visiting adults and their larvae. Maggots have been reported inside two other species' inflorescences. Spiders, cockroaches and ants turn up in every one of the more detailed studies. The idea is old: it was argued in 1889 about a related aroid, that the flies were the pollinators and the beetles merely followed them in to eat them.

Nobody has investigated it in this genus. It is listed in the sources block as one of the open questions, because a grower who sits up with a flowering plant and a notebook is in a position to say something about it.

What was actually found inside a wild titan arum

The visitor records above are almost all from cultivated plants. Three visitor records for the titan arum come from wild plants; the most detailed is from Sumatran forest, where a flowering Amorphophallus titanum was opened and its chamber emptied, and what came out was a beetle assemblage rather than a single partner: Histeridae and Hybosoridae — including Phaeochroops emarginatus, the same genus recorded from Amorphophallus paeoniifolius — plus weevils, rove beetles, scarabs and cockroaches. Stingless bees, Trigona geissleri, were photographed resting on the stigmas.

The same observer recorded something else worth carrying: at about five in the afternoon the smell changed character, from rotting flesh to excrement. Two audiences, or one audience and a shift in what the plant is claiming to be.

A fifth species, watched for eight years

The most recent addition to this thin list is Amorphophallus napalensis, in Nagaland, north-east India, followed at a wild site and a cultivated one from 2008 to 2016. The visitor is a scarab. The study calls it Parastasia; three beetle specialists who went back to its own photographs read it as Peltonotus — a genus already tied to this one, on Amorphophallus paeoniifolius in Thailand. Either way the account is unusually specific about what the beetle does rather than merely that it was there. It arrives at dusk — the observer records that nothing else comes at night — shelters through the following day at the base of the spadix under the in-rolled spathe, and leaves around half past five in the afternoon, flying up to fifteen meters at a stretch to the next inflorescence that has just opened and is smelling strongly.

Three other insects were watched and explicitly set aside, which is rarer and more useful than it sounds: honeybees and stingless bees arrive only on the third day and were never seen to reach the spadix base, and the drosophilids stayed on the male flowers and the appendix and were “never found visiting female flowers.”

Wild and cultivated sites, Nagaland · eight seasons · no sample size stated anywhere in the paper · the only figure it reports is the fifteen-meter flight

A second kind of payment

Part of this genus feeds its beetles on sterile flowers. Another part does something different: in Amorphophallus galbra and Amorphophallus variabilis the inside of the spathe base carries an area of greasy yellow tissue, and those species smell of fermenting fruit rather than carrion.

For Amorphophallus variabilis that tissue has actually been sectioned, by the man who first described it in 1937. The yellow patch runs a centimeter or two up from the spathe floor, and beneath it lie about ten layers of cells packed with starch grains and oil droplets — the outermost two almost pure oil, with very delicate walls. He cut a spathe off, put it under a microscope with the beetles still on it, and watched their mouthparts working at the surface. After a few days of feeding, white patches mark where they have eaten down to the tissue beneath. The plant is not damaged; it has grown a layer to be eaten.

So there are at least two independent reward systems in the genus — edible sterile flowers, and an edible patch on the spathe floor — and they go with different smells.

Months ahead: is it going to flower at all?

After dormancy, the bud at the center of the corm becomes either a leaf or an inflorescence — one or the other, not both. And you can tell which, by hand, well before it commits to anything visible.

Nursery observation · Amorphophallus paeoniifolius, 46 cultivated inflorescences at Bogor, followed daily Measured

A glasshouse protocol for Amorphophallus titanum makes the same split by shape rather than by feel, and it adds a cue the squeeze test does not have. A leaf bud is arrow-shaped and perfectly round in diameter, with the tip of the cataphyll exactly in the center. An inflorescence bud is bell-shaped, irregularly rounded, and its cataphyll tip sits clearly off to one side — which, the authors write, makes the bud look “pregnant”. They call it readable by an experienced eye at about 60 mm of bud diameter, and obvious once the bud stands more than 100 mm above soil level.

Long-running glasshouse practice · Amorphophallus titanum · a shape description with two rules of thumb — no measurement series, no sample size and no error rate are printed Practice

Weeks ahead: the stages

From a visible bud to full bloom takes roughly three to five weeks in the species this has been measured in. The sequence is reliable even where the day counts are not.

It can also be startlingly fast. A wild Amorphophallus gigas followed in Sumatra went from a 58 cm bud to 166 cm in four days — better than 25 cm a day — then slowed, reaching its full 230 cm two days before the spathe opened. Height finishes first; the spathe opens on a stem that has already stopped growing.

Amorphophallus gigas, North Padang Lawas, Sumatra · one inflorescence, measured every two days Single plant

A cultivated plant of the same species sits inside the three-to-five-week range above. At Fairchild Tropical Garden in Miami, an Amorphophallus gigas inflorescence came up around May 1st, 1999 and began releasing a powerful odor on May 28th, about four weeks later, from a tuber that weighed only 4.5 pounds (about 2 kg).⁷²,⁷³ Its keeper, Craig M. Allen, took the emerging shoot for a new leaf at first, and noted that neither that plant nor the second Amorphophallus gigas to flower that spring had the tight bract around the emerging inflorescence that was so distinctive in the titan arum, so the look of a titan arum bud should not be assumed on other species.

Stage timings · Amorphophallus paeoniifolius, 46 inflorescences, 10 measured in detail Measured   A formal growth-stage scale for this species was published in 2026 and is cited in the sources.

On the day: the three checks

1 · Is it fully open?

The limb has to have finished unfurling. A spathe that is still parting is not yet in the female phase, and cutting into it early wastes the window you were trying to protect.

2 · Does it smell?

Full-strength scent is the female phase. Whatever your species smells of — and Part V is the reason not to assume it is carrion — peak intensity and receptivity coincide. In practice the field literature uses the smell as the trigger: pollinate after the scent has developed.

One published hand-pollination program in the genus put it exactly that way: “Where possible we pollinated after the development of the scent.”

3 · Are the stigmas wet?

This is the check that decides it. A receptive stigma carries visible fluid and is tacky to a brush. The published descriptions say the stigmas are “very sticky and no pollen will fall off” — which is both the test and the reassurance. If the surface is dry, you are early or late.

Two more free instruments

The appendix changes size. In four dated photographs of one anthesis, it was larger as it approached peak heat and then slightly smaller than its starting size around the time the pollen was released. A shrinking appendix may therefore go with the arrival of the male phase, though this rests on one plant and nothing was measured. If you are waiting to collect pollen, it is a cue to watch for.

And it changes color. Through the same sequence the appendix went pale green to pale reddish-purple, the lower inner wall of the spathe went red to deep purple, and the upper inner wall went bright green to bright reddish-purple. The author's suggestion is that the shift is a visual signal to insects; either way it is a maturity clock you can read without touching the plant.

A free thermometer

If you want to know whether your plant is heating and you have no instrument, look at the surface of the spadix. One 2020 study of 46 nursery-grown Amorphophallus paeoniifolius inflorescences used droplet density as its heat proxy for exactly this reason — the more droplets standing on the appendix, the hotter it is running. That was the study's stated assumption, not its finding. It never put a thermometer, a thermocouple or a camera on the plant, and nobody has since checked the cue against a temperature. Read the droplets as a sign that anthesis is under way, not as a reading of how hot the plant is.

Look at the inside wall of the spathe too. In one plant imaged with a thermal camera on the day of peak heat, that wall was the coldest surface in frame — a few degrees below the room — while the appendix above it ran about five degrees over the room, and the wall was reported moist and beaded with droplets that day. Hot column, cold wet wall, strong smell, all at once.

What those droplets are is disputed. One reading is condensed water, which would mean the plant is losing most of its heat to evaporation. The other is that they are the scent secretion itself, arriving at the surface. Part IV has the argument. For a grower it does not matter much: either way, droplets mean the performance is at full tilt.

If the stigmas are wet and you have pollen: go to Part IX. If they are wet and you do not: there is nothing to be done today, but Part VIII is how you make sure that is not true next time.

Read the storage results the right way round

Here is the run of storage results from that program, and the useful thing about it is that it does not sort by duration.

One private collection, indoors, no insects, no wind · 20 pollination events across 1999–2003 One program

Stored against fresh, on one spadix

A separate program put the comparison on a single plant. In May 1996 the spadix of a cultivated Amorphophallus titanum at the Bonn botanic garden was divided into four sectors, each pollinated with differently stored pollen — the garden's own stored material against pollen harvested fresh at the Palmengarten in Frankfurt. The reported outcome was “no difference in the number of fruits, the contents of seeds in each fruit or in the germination rate between the treatments”.

Read that the way the table above has to be read. No per-sector counts and no statistics were printed: the conclusion is stated rather than tabulated, so it means no difference was apparent on one inflorescence, not that a difference was looked for and excluded.

So what actually explains the failures?

Probably not time. The growers who ran the storage program above identify their own likeliest confound and it is a mundane one: they used no desiccant, and “in some cases we found the pollen wet after opening the tubes”. Their stated intention for next time was to “dry the air in the tubes before freezing”.

That lines up with what the Alocasia guide concluded from entirely separate sources: what ruins stored pollen is trapped damp. Two genera, two literatures, one answer. Until somebody runs the assay this genus is missing, treat moisture as the variable you control and duration as the one you tolerate.

Collecting it

The unripe-pollen route, and its one honest test

A botanic garden reported selfing Amorphophallus titanum using pollen dissected out of anthers that had not yet opened — which, if it works, defeats protogyny entirely, because you can take pollen from an inflorescence on the same day its own stigmas are receptive.

It was tried on Amorphophallus beccarii, with a magnifying glass, a scalpel and needles. It failed. Four ovaries swelled and then the peduncle withered. The growers' own explanation is size: Amorphophallus beccarii anthers are far smaller than a titan arum's, and dissecting immature pollen cleanly out of them may simply not be possible by hand.

So: a real technique on a large-anthered species, unproven and once-failed on a small one. If you have a titan arum and no donor, it is the technique to try on the inflorescence that is open now; another same-plant route, banking this flowering's pollen for the plant's next one, is in Part IX. If you have something the size of a konjac, do not count on it.

With pollen in hand and a wet stigma in front of you, go to Part IX.

Gloves, and why specifically

The usual aroid caution applies here with an unusually direct mechanism. Anatomical work on Amorphophallus titanum found calcium-oxalate raphides — needle-shaped crystals 30–90 µm long, grouped in bundles in the spathe — through the spathe and the ovaries, and, pointedly, raphides lying free in the stigmatic fluid, as seen in an electron micrograph of fixed tissue.

That is the fluid you are about to put a brush into and then, at some point in the next ten minutes, touch your face with. Gloves, and keep the brush away from your eyes.

Protocol from one published hand-pollination program in the genus · private collection, 24 inflorescences, 20 pollination events, 1999–2003 One program

Keeping pollen until the other plant opens

Two inflorescences almost never open together, so most crosses in this genus are made with stored pollen. At room temperature viability falls away within about two days. Cold buys a great deal more.

The two documented grower protocols work the same way — dry it first, then hold it cold. Put the pollen in an open tube inside a sealed jar with silica, leave it in a fridge at 4–7 °C for one to four days to dry, then cap the tube and move it to a freezer at −15 to −20 °C. The lighter version skips the freezer: a paper envelope laid flat in the fridge with a silica packet on top. Thaw or warm for ten minutes before use, and apply it immediately.

A botanic-garden protocol for Amorphophallus titanum describes no drying step at all — the pollen goes either into liquid nitrogen or into a refrigerator at 5 °C — so dry-then-cold is how the grower protocols work, not how every documented one does.

Four species across five storage regimes, and different definitions of “still good” — a cross that set seed or fruit, or only a report that the pollen worked Not comparable

Which is the honest caveat: these numbers are not a ranking. Nobody has run one species through all of them. And pollen consistency differs across the genus — dry in Amorphophallus konjac, wetter in Amorphophallus paeoniifolius — so a protocol that suits one may not suit the next. Treat the table as four separate demonstrations that cold works, not as a league table.

Can you self it?

That depends entirely on the species, and the two ends of the range have now both been tested properly. One Amorphophallus refuses its own pollen outright. Another accepts it and fills the whole spike. Between them sits a scatter of grower reports, below.

So the genus holds both answers at once, and neither generalizes. Any sentence beginning “Amorphophallus is self…” is wrong before it finishes. What follows is the middle ground — mostly grower reports, one or two plants at a time, with one published experiment among them.

The pattern, such as it is: a self needs a second inflorescence to carry it — another bloom on the same plant, or an offset, opening at the right distance in time. On one inflorescence alone, protogyny usually wins. What a successful self then yields is not predictable from the fact of it: three viable seeds in Amorphophallus curvistylis, 24 berries in Amorphophallus eichleri, fruit across the whole spike in Amorphophallus titanum and again in Amorphophallus decus-silvae — that last one photographed rather than counted. None of it is genetically new.

Set beside the two measured cases above, the reading changes slightly. A poor self is not proof of incompatibility — in most of these the timing is the obstacle, and three viable seeds is what beating the timing badly looks like. Only Amorphophallus albus has been shown to refuse pollen it could physically receive.

Crossing between species

It has been done. Two interspecific hybrids were raised at a European botanic garden — Amorphophallus longituberosus × Amorphophallus albispathus, and Amorphophallus dunnii (published as Amorphophallus odoratus) × Amorphophallus yunnanensis — though neither was released. Cross-clone pollination within a species is described as “usually successful”. One named case: the Amorphophallus lewallei seed lot that Wilbert Hetterscheid distributed was obtained by crossing two of his clones, H.AM.020 and H.AM.021, which differ in the warts of the appendix, conical and pointed in the first and rounded in the second.⁷⁴

Against that, an attempted cross between two forms of the same species, Amorphophallus beccarii, produced only slight ovary swelling and then nothing. Compatibility here is not a solved question, and there is still no matrix for it — but there is now a substantial run of attempts, kept by the growers making them, and a rough rule falls out of it.

Close relatives usually take. Amorphophallus puttii × Amorphophallus thaiensis and Amorphophallus saraburensis × Amorphophallus scutatus both worked. Distant ones usually do not — Amorphophallus borneensis × Amorphophallus longispathaceus (recorded as Amorphophallus dactylifer), Amorphophallus lewallei × Amorphophallus bulbifer, Amorphophallus myosuroides × Amorphophallus lunatus.

And then the rule breaks, usefully. Amorphophallus variabilis × Amorphophallus titanum has now succeeded twice, between two plants nobody would put side by side: once in 2002, and again in 2017 from a second Amorphophallus variabilis mother, that second cross flowering as Amorphophallus ‘Digest’ in December 2019. Amorphophallus aphyllus × Amorphophallus consimilis and Amorphophallus henryi × Amorphophallus yunnanensis did the same.

Most of the hybrids grow faster and heavier than either parent. One Amorphophallus albus × Amorphophallus konjac tuber reached 1,200 g and 14 cm across in three growing cycles; an Amorphophallus lewallei × Amorphophallus maximus reached 242 g in two. Amorphophallus aphyllus, slow and tender on its own roots, produced a hybrid that offsets freely and flowered in its third season.

One more thing worth knowing before you conclude anything from a failure: a cross that fails once has not been ruled out. Pollen age, the length of the female window and the exact hour all confound it, and the growers keeping these records treat a single failure as no result at all.

Every recorded cross

The growers keeping these records have now logged more than five hundred attempts, and the failures are kept as carefully as the successes — a pairing that has failed four times with frozen pollen is worth knowing about before you spend an inflorescence on it. Search either parent, or both. A cross that took links to its hybrid’s own entry; a bracketed parent is itself a hybrid.

Pollination worked if the ovaries keep growing after the rest of the inflorescence has gone. In Amorphophallus paeoniifolius, measured: the shine leaves the spathe a day after full bloom, the whole structure withers at three to five days, collapses at eight to eleven, and the peduncle detaches from the corm at thirteen. None of that tells you anything about the seed.

The berries then take months, and how many depends on the species and the program. One hand-pollination program recorded two to four months from pollination to harvest. A cultivated Amorphophallus titanum hand-pollinated at the Bonn botanic garden in May 1996 took seven to eight months, harvested that December and the following January — one infructescence, dated at both ends.

What is inside

Not much, per berry, and the best data come from a field study that opened several hundred of them.

Wild and cultivated material, Uttara Kannada, India · failed and undeveloped fruits excluded from the counts Measured

The hand-pollination program's own harvests agree in shape if not in detail: “the berries contained either one or two seeds, and occasionally no seed at all.” Seed size varies enormously with species — from about 80 mg to over 750 mg in that one collection.

Cultivated Amorphophallus titanum fits the same shape. In 1999, hand pollination at Selby Gardens⁴⁹ was followed by about 200 berries, though not all held seed: about 80 seeds in all, an approximate count because a few berries were given out early and never tallied. Some berries held two seeds, many held one, and most of the seeds germinated, even the very small ones.⁵⁰

The ceiling has an anatomical reason. In the species where the ovary has been described in detail it carries two cells with one ovule in each, so two seeds is the structural maximum and anything less is an ovule that was not fertilized. Species with three- and four-celled ovaries exist, which is where the occasional third and fourth seed comes from.

Amorphophallus decus-silvae sits in that second group. A hand-pollination experiment on it photographed berries holding one, two or three seeds, which puts at least three locules in its ovary if each locule carries a single ovule. That count is arithmetic across two descriptions rather than a dissection: the published account of this species’ ovary gives its size and its stigma but not how many cells it holds. The number of fruits opened is not stated either — so read that as a described range, not a distribution.

One oddity worth knowing about, because it will look like failure: fully developed berries collected from wild Amorphophallus prainii in Malaysia contained no seed at all — only fragments of seed coat. Nobody has explained it.

Sowing

Germination percentages for the grower program are derived here from its published seeds-sown and seedlings-raised columns; that source tabulated counts, not rates. The Amorphophallus titanum figure is printed as a rate in its own paper. Two programs

What a successful cross costs the plant

This is the part nobody mentions until afterwards. Among the Asian non-evergreen species, a plant that has flowered and been effectively pollinated will not produce a leaf that season. In the wild, fruiting plants are simply never found carrying leaves. African species behave differently and leaf out regardless of whether they set fruit.

The word effectively is load-bearing, and a separate Indian field study shows why. There, flowering shoots that decayed without setting fruit were followed by fresh leaves from the same tubers, in every case, wild and cultivated alike. So it is not flowering that costs the plant its season — it is fruiting. A cross that fails leaves the corm free to make its leaf and try again next year.

Where the seed goes in the wild

Birds — and unlike almost everything else on this page, this part has actually been watched rather than inferred. In the Western Ghats, two bulbuls and the koel are the regular takers, well enough known locally that people snare birds at fruiting plants. A captive koel offered ripe fruit ate hundreds over two days, and both regurgitated and defecated the seed — regurgitation about 35 minutes after feeding, defecation later. Since the gap decides how far a seed travels, the two routes disperse very differently.

Gape size sorts out who gets what. The lower, larger fruits on a robust Amorphophallus paeoniifolius spike are too big for a bulbul to swallow; when only those are left, bulbuls stop visiting and the koel finishes the job. Hornbills are the usual answer for the fruits of Amorphophallus titanum in Sumatra — berries that the two wild West Sumatran infructescences anyone has measured, one per site, put at 0.8–4.1 cm and 0.7–3.9 cm long. It is worth knowing how thin that answer is. It began as an inference from the size and color of the berries; a survey of fruiting plants in West Sumatra saw no hornbill visit one; the published diet lists for all six Sumatran hornbill species do not include it; and hornbills feed in the high canopy, while an Amorphophallus titanum infructescence stands on the forest floor. The same survey found the plants growing in tight clusters — the pattern you get when seed rolls and washes a short distance, not when a bird carries it. Birds of paradise have been reported on Amorphophallus paeoniifolius in New Guinea. Whether a species' fruit gets moved at all seems to depend on how tidily it is arranged on the spadix: Amorphophallus bulbifer, whose fruits sit irregularly on a thin watery peduncle that bends when a bird lands, is dispersed noticeably less well than its neighbors.

Part XI is the corm the seedling is now building, and why it will be several years before any of this happens again.

These are geophytes: the plant is a corm, and everything above ground is temporary. In a growing season the corm sends up one leaf — a single dissected blade on a long petiole, umbrella-shaped, sometimes enormous. The leaf feeds the corm, dies back, and the corm sits dormant. That is the usual pattern rather than the only one: Wilbert Hetterscheid described Amorphophallus coaetaneus as an evergreen species whose leaves may stay on for three to five years, and advised keeping it in growth and refreshing the soil every two years.⁷⁷

Then, at the start of the next cycle, the bud at the center of the corm becomes either another leaf or an inflorescence. Not both. Part VII is how to tell which, by feel, months in advance.

Young seedlings do not always keep to a single leaf. Ron McHatton, raising about 4,000 Amorphophallus titanum seedlings in Apopka, Florida, found that virtually all made a series of leaves in rapid succession, as many as five, before the first dormant period. The few (perhaps 1%) that did not follow that pattern, and that he dug up, all had a problem with the roots.⁵¹ Other growers reported less. One had Amorphophallus titanum seedlings go dormant after a single leaf,⁵² and another, citing a tissue-cultured plant, called up to three leaves before the first dormancy quite normal.⁵³ Another species ran a long time before its first rest: Allan, a grower in Canada, raised about 30 Amorphophallus hewittii from seed, lost only a couple, and saw the seedlings keep making leaves for about two years before going dormant for the first time, with dormant periods of only about two to three months.⁷⁸ Ron McHatton found soil temperature important, at least in juvenile Amorphophallus titanum, for keeping them growing: the optimum appeared to be near 80 °F (about 27 °C), the plants tended to overheat if it went much higher, and wet potting mix can be as cool as 50 °F (10 °C) through evaporation when the air is 60 °F (about 16 °C).⁷⁹

The cycle, as it has been formally described

A growth-stage scale was published for Amorphophallus paeoniifolius in 2026, from a year of fieldwork in the Philippines — the first formal phenological framework for the species. It runs from dormancy, through leaf development and pseudostem elongation, into inflorescence and fruit development, and back through senescence to dormancy. In that population the cycle was bimodal and monsoon-locked: a long dormant phase over the drier months and an active reproductive-and-vegetative phase through the wet ones.

That is the frame to hold. The practical numbers below sit inside it.

The same work puts numbers on the end of the sequence that the rest of this page could not: once the spathe is fully open, anthesis begins within 24 to 48 hours and lasts up to two days; the spathe and appendix are withered three to seven days after it ends; and if pollination worked, fruit takes 150 to 180 days to develop. In that population flowering was confined to June and July, and about a quarter of the plants were reproductive in the survey year, the rest in leaf.

Amorphophallus paeoniifolius · Bohol, Philippines · 1,264 individuals followed · one year, June 2024 to May 2025 · wild and cultivated mixed Measured

In a replicated trial of plants grown as Amorphophallus muelleri (n = 48 corms per depth), three-year corms planted at 5 cm produced a mean 185.3 seed pods per inflorescence, compared with 163.7 from corms planted at the surface. The same trial (n = 48 corms per age treatment) found 96% of three-year corms flowering, versus 8% of two-year corms and none of the other age treatments.

Two further intervals come from one grower in the wet tropics. At Whyanbeel Arboretum in Far North Queensland, Michael Pascall raised Amorphophallus henryi from seed received from Wilbert Hetterscheid in 1996, and it grew, flowered and set seed within four years, though he allowed that the tropical climate may have helped.⁸⁴ He also reported that an Amorphophallus paeoniifolius tuber of more than 30 cm flowered about two years after starting as a 2 cm tuber, growing outdoors in ideal wet tropical conditions with plenty of mulch and fowl manure; his report gives no tuber weight and no planting date.⁸⁵

One thing that can go wrong before anthesis

The rule that a corm makes either a leaf or a flower is a tendency, not a law, and when it breaks the result is bad. In a 300-tuber trial of Amorphophallus muelleri, a leaf and an inflorescence emerged together on 57 tubers — about a sixth to a fifth of them. The authors proposed that the emerging petiole strangles the peduncle and that the dense mat of roots at the petiole base adds pressure.

Can you make it flower?

There is one hormone trial in the literature and it deserves careful reading, because it is easy to make it say more than it does. Three hundred tubers of Amorphophallus muelleri, gibberellic acid and an auxin, five concentrations each, sprayed onto the bud and planted the next day.

300 tubers · Amorphophallus muelleri, 3 years old, >500 g, buds already visible · conference proceedings, peer-reviewed Measured

One more thing about a flowering tuber

It barely roots. During inflorescence growth Amorphophallus muelleri tubers produced only five to eight roots — a known trait of the species while flowering and fruiting. A plant in that state is not taking up much of anything, which is another argument for feeding it in the previous season. Heavy feeding of another species has a dated report behind it: Alistair Hay of the Royal Botanic Gardens Sydney wrote on October 9th, 2004 that the horticultural staff had deliberately overfed a subset of the garden’s Amorphophallus titanum plants; the pots toppled over as the tubers swelled, and one plant, at 15.5 kg, produced a flower. It is one flowering with no untreated comparison, and the post does not say when the feeding was done.⁸⁶

Why the plant does not really need you

Worth facing directly, because it explains a great deal about the state of this genus in cultivation: Amorphophallus has several separate ways of copying itself without any sex at all — though not every species has all of them.

None of that is universal. For Amorphophallus titanum, a long-running glasshouse protocol for the species reports “no indications that the Titan Arum self-propagates in this way” — no bulbils, no rhizome-like offsets. Where that species has been increased vegetatively it has been by hand, from leaf cuttings, not by anything the plant does on its own.

One conservatory has described that hand route in detail. At the University of California, Davis Conservatory, Amorphophallus titanum has been propagated from leaflets by a method developed there by Ernesto Sandoval. D. Christopher Rogers described it as terminal leaflets with a petiole about as thick as an index finger, the lamina stripped from the part of the petiole to be buried, set upright in vermiculite and perlite, treated with a low dose of rooting hormone, and given plenty of bottom heat and high humidity.⁵⁵ He reported a small bulb in about two months and roots about two months after that; Amorphophallus titanum leaflet cuttings with a petiole about 2 cm thick or more were about 80% successful, against about 25% for thinner petioles. Younger leaflets worked best, and leaflets from older leaves never worked.⁵⁶ No sample size was given.

The negative is narrower than it sounds. Wilbert Hetterscheid wrote that Amorphophallus titanum does not produce offsets, which he defined as true side tubers or rhizomes that develop on the mother tuber and separate from it by themselves in their first year or at the start of their second; the phenomenon another poster had described, he wrote, was the development of adventitious buds into new tubers, which he called splitting.⁸⁷ Doug Ewing, greenhouse manager at the University of Washington, reported that he had grown Amorphophallus titanum plants that did offset: they did not multiply like potatoes, but seemed now and then to yield an extra tuber or two.⁸⁸

The other routes in the table above are not uniform either. Wilbert Hetterscheid wrote that some clones of Amorphophallus coaetaneus develop bulbils in the main leaf axis and others do not, and advised leaving them on the leaf until it has died back completely before planting them.⁸⁹ Amorphophallus longispathaceus, grown as Amorphophallus dactylifer, is reported to make no bulbils on the leaf: Ken Mosher never had a plant of that name produce one, found instead lots of very small offsets at the top of the tuber that must be left attached, and thought a plant of that name that does bear a bulbil was probably misidentified, and Hetterscheid agreed that the true plant has no foliar bulbils.⁹⁰,⁹¹

Why this table is shorter on causes than you might expect

Because the experiments that would separate them have not been done. There is no run of viability counts over storage time for Amorphophallus pollen, no measured receptivity window for any species but the cultivated elephant foot yam (Amorphophallus paeoniifolius), no quick test for stigma receptivity, and no compatibility matrix for the genus. One published program of repeated attempts concluded that its failures came from its own varying conditions and then added that it could no longer verify that assumption.

That is the honest position, and it is why the middle column of this table exists. When a cross fails here, you usually cannot know why — and a troubleshooting guide that pretends otherwise would just be inventing diagnoses.

Peer-reviewed literature, one preprint, and grower reports

Listed with the reason they were not used

Eleven things nobody has measured

Each of these is a statement this page cannot make, because the measurement does not exist. They are listed because several are genuinely within reach of a grower with a microscope, a notebook and patience — which is not true of most research gaps.

Colophon

Contributor Written by Robert Wainblat for Aroidpedia, July 2026. Part of the Aroid Pollination series.