Short answer
A prayer plant (Maranta leuconeura) sets seed rarely in cultivation, and each flower can make at most three seeds, because the ovary has one ovule per chamber and often only one fertile chamber [1]. Seeds in the family carry a fatty appendage, the aril, that ants harvest [2][4]. In a study of a Central American relative, ant-handled seeds sown into steady moisture germinated readily, while delayed seeds became dormant [4]. No published germination protocol for Maranta leuconeura itself was located.
Why does a prayer plant make so few seeds?
The limit is built into the flower. The Flora of North America family treatment describes the Marantaceae ovary as inferior and three-chambered, with one to three of those chambers fertile and a single ovule in each [1]. An ovule is the structure that becomes a seed after fertilisation, so one Maranta flower can yield at most three seeds and often only one. The Maranta clade, the group that contains Maranta, Ctenanthe, Stromanthe and their relatives, is characterised by fruits that are dehiscent capsules, meaning they split open when ripe [7].
Fruit set is low even in the wild. Ley and Claßen-Bockhoff tested 28 African Marantaceae species in Gabon: all were self-compatible, yet only two set seed on their own, and natural fruit set averaged about 11% [6]. Hand self-pollination raised fruit set considerably in every species [6]. The reason is the family's explosive flower: pollen is presented on the back of the style, which is held under tension and only springs when a visitor trips a trigger, so a flower that is never visited rarely pollinates itself [10]. Indoors there are no visiting bees, which is the main reason a flowering prayer plant usually drops its flowers without fruit.
Some cultivated forms do not flower at all. The patent for Maranta leuconeura 'Beauty Kim' states that flowers had not been observed on the new cultivar [9], and named forms like it are kept true by cuttings and tissue culture, not seed.
What is the aril on a Maranta seed for?
The Flora of North America lists an aril on the seeds of nearly all Marantaceae [1]. An aril is an extra outgrowth from the seed stalk that partly wraps the seed. Hammel's account of the family in Costa Rica describes the seeds as often blue-grey to black and provided with an oily aril [2], and he reads the oily aril of the inconspicuous Ischnosiphon fruits as a sign of dispersal by animals [2].
In several neotropical Marantaceae the animals are ants, a dispersal system called myrmecochory. Horvitz studied seeds of the plant she named Calathea microcephala carried by three genera of neotropical ants: the ants moved seeds rapidly to their nests, chose where the seeds ended up, removed the arils for food and did not eat the seeds; the arils were rich in lipids [4]. A global analysis of myrmecochory by Lengyel and colleagues names Calathea among large genera in which ant dispersal is well known [11].
Whether Maranta leuconeura itself is ant-dispersed in its Brazilian forests has not been studied; no field record was located as of 2026-10-09. The aril is there, and the Flora e Funga do Brasil records the species as a ground-dwelling herb of gallery, semideciduous and rain forest [8], the same forest-floor setting in which ant dispersal has been documented in the family.
What does ant handling do to germination?
Horvitz's experiments in captive ant colonies give the most direct evidence on how Marantaceae seeds behave once shed. Seeds taken immediately to a consistently moist spot germinated readily, 72% of them, with a mean germination time of 29 days, and for those seeds removing the aril made no significant difference [4]. Seeds that waited before reaching suitable conditions behaved differently: they seemed to enter an induced dormancy, germinated less often and took up to 85 days even after conditions improved, and for them aril removal enhanced germination [4].
Later field experiments with Calathea ovandensis in Mexico extended the picture. Seeds that carried their ant reward produced 1.6 times as many seedlings as seeds without it, and the seedlings emerged farther from the parent [5]. Light mattered even more: in germination boxes, seedling emergence in treefall gaps was about 180 times that in deep shade, and the ungerminated seeds in deep shade were mostly dormant rather than dead [5]. The authors describe a relatively long-lived seed bank and note the unusual feature that the seed body stays attached to the new seedling [5].
What is inside the seed?
The food store of a Marantaceae seed is not the usual endosperm. The Flora of North America describes endosperm as scanty or absent, perisperm as copious and the embryo as curved, with one or two canals running through the perisperm [1]. Perisperm is a storage tissue derived from the ovule's own nucellus rather than from fertilisation, so the embryo feeds on maternal tissue as it germinates.
In a comparative study of seeds across the ginger and banana order (Zingiberales), Benedict and colleagues examined 23 species in 20 Marantaceae genera and found the perisperm canal to be unique to the family; the seeds are also operculate [3]. An operculum is a small lid of seed-coat tissue over the point where the root emerges; the embryo pushes it off to germinate. Within the Maranta clade the perisperm canal is branched at its base [3]. Whether the operculum or the seed coat controls the induced dormancy Horvitz measured in delayed seed [4] has not been tested, so the lid is best read as the point to keep moist, not as a barrier to break.
Sowing prayer plant seed: what the biology suggests
No germination protocol for Maranta leuconeura was located in Europe PMC, the Flora e Funga do Brasil or Google Patents as of 2026-10-09. What follows is drawn from the family studies above and should be read as such.
- Get seed by hand-pollination. With no bees indoors, seed on a home plant almost always means a grower tripped the flower's trigger and moved pollen; hand self-pollination raised fruit set in every species Ley and Claßen-Bockhoff tested [6].
- Sow fresh. Seeds that met moist conditions at once germinated at 72% with a mean of 29 days, while delayed seeds slid into dormancy [4].
- Remove the aril. It made no difference to fresh seed but helped seed that had become dormant [4], and a fatty aril left in a warm, wet tray invites mould.
- Keep the surface steadily moist and warm, the condition of an ant nest chamber in forest soil [4].
- Give light, not darkness. Gap light raised seedling emergence about 180-fold over deep shade in Calathea ovandensis [5]; bright shade, never direct sun, is the indoor equivalent.
- Be patient. Up to 85 days is within the range Horvitz recorded [4].
Seedlings are not copies of the parent. The leaf colours of forms such as 'Kerchoveana', 'Erythroneura' and 'Lemon Lime' are kept by vegetative propagation, and the 'Beauty Kim' patent records that cuttings and tissue culture reproduced that form true to type [9]. Seed from a hand-pollinated plant may segregate into different leaf patterns, which is the main reason to sow it.
Frequently asked questions
How many seeds does a prayer plant flower make?
At most three. The Marantaceae ovary has three chambers, one to three of them fertile, with one ovule in each [1]. Many Maranta flowers produce a single seed or none, since fruit set without a pollinator is low across the family [6].
Why are there no seeds on my flowering prayer plant?
Marantaceae flowers are self-compatible but need a visitor to trip the style. In Gabon only 2 of 28 tested species set seed on their own, and hand self-pollination raised fruit set in all of them [6]. Indoors, without bees, flowers usually fall unfertilised.
Should the aril be removed before sowing?
It helps with seed that has been stored. In Horvitz's experiments aril removal did not change germination of seed sown at once into moist conditions, but enhanced germination of delayed, dormant seed [4].
Will prayer plant seedlings look like the parent?
Not necessarily. Named leaf forms are kept true by vegetative propagation; the 'Beauty Kim' patent records cuttings and tissue culture reproducing that form true to type [9]. Seedlings are new genetic individuals and can differ in leaf pattern.






