Ants and a New Look at Erythronium Part 1

Revised January, 2011.  Copyright 2011 Arthur G. Guppy

An earlier version of Part 1 was published in the Summer 2007 issue of the Rock Garden Quarterly with photos of Erythronium japonicum, E. sibiricum (unmottled leaves), E. californicum with red nectar guides, E. revolutum, E. tuolumnense, and seeds of E. japonicum showing elaiosomes


The genus Erythronium divides naturally into two subgenera,  one that is myrmecochorous [seeds are dispersed by ants] and one that is not.  The two subgenera are geographically widely separated and behave very differently both in nature and in the garden.  

The association in the title of Erythronium with ants may seem strange, but there is a connection that those who grow Erythronium need to be aware of, especially if they raise the plants from seed.

I discovered the association by chance many years ago when I obtained Erythronium japonicum seeds from a seed exchange and noticed little projections on the seeds.  Until then my experience with seeds of the genus had been limited to species from western North America, and I was sure that none of those had projections on the seeds.  Fortunately I had been reading E. I. Applegate’s monograph on the Erythronium of western North America (1935), and I had also been reading an article on myrmecochory, which is the dispersal of plant seeds by ants.  Applegate had written that the Erythronium of Europe, Asia, and eastern North America comprise an entirely different group from the Erythronium of western North America.  (Those two groups within the Erythronium of the world mentioned by Applegate should not be confused with his well-known division of the Erythronium of western North America into two sections, the Pardalinae and the Concolorae, on the basis or whether or not the leaves are mottled.  I will come back to those sections.)  The fortunate juxtaposition in my mind of Applegate’s grouping of Erythronium with seed dispersal by ants made me wonder if the projections on the seeds of a species from Asia could be elaiosomes.


Elaiosomes are little fleshy projections on plant seeds that are irresistibly attractive to ants.  When ants discover such seeds, they seize them and carry them away to their nests, where presumably the elaiosomes are fed to the ant larvae.  The seeds are then discarded, much as we discard cherry pits, though with many ant species the discarded seeds are carefully removed from the nest and dropped some distance away.  With a little luck, some of the seeds are dropped where they can grow into new plants, well away from where they would compete with the parent plants.  Myrmecochory is the name given to this extraordinary natural process by which plants make use of the remarkable energy of ants to get their seeds dispersed to new growing sites.  It is such an effective way of dispersing seeds that it is known to occur in over 3,000 plant species and has been observed on every continent except Antarctica.  Some well-known genera with myrmecochorous species are Claytonia, Corydalis, Dicentra, Euphorbia, Fremontodendron, Trillium, Vancouveria, Viola. 

Almost certainly this use of ant energy evolved separately in many different plant families, and consequently the elaiosomes of different families, and even different genera and species, may contain some different attractive substances, though basically they are generally lipids.  It is evident that in some cases, in addition to providing food for the ants, or possibly instead of providing food, the elaiosomes contain substances that work directly on an ant’s instinctive urge to carry an object to its nest. For example, oleic acid, which is present in the elaiosomes of at least some Trillium species, is believed to trigger the corpse-carrying instinct in ants (Lanza et al. 1992).

Though there is great variation between genera and even species, generally elaiosomes emit a volatile attractive scent which is quite quickly dissipated after the seeds are exposed to air by the dehiscence of the seed capsules.  This rather quick loss of attractive power might seem disadvantageous, but there is a way it can improve the chances for successful dispersal of the seeds.  One of the purposes of myrmecochory is to avoid the accumulation of large numbers of seeds under the parent plant, as the numerous seeds would attract birds or other seed predators.  However, the accumulation of discarded seeds in a midden near an ants’ nest would be even more likely to attract predators.  I have frequently watched an ant pick up a seed and set off with it towards its nest, only to lose interest in its burden, drop it, and walk away.  It seems evident the scent from the elaiosome had given out and left the ant with no motivation to continue carrying the seed.  Other observers also have observed ants carrying seeds for only a short distance before abandoning them.  There is an obvious advantage for the seeds from this behavior for they avoid being in an accumulation either under the parent plant or in the ants’ midden.

Not all ants are helpful dispersers of seeds.  There is a report from Australia of an ant genus that frequently consumes elaiosomes in situ, leaving the seeds where they had fallen under the parent plant (Anderson & Morrison 1998).  In warm, fairly dry regions there are harvester ants that live by collecting seeds to use as food.  There are many species, and they tend to be large, fast-moving ants that range over large areas.  Their activities tend to defeat the purpose of myrmecochory, as they consume both elaiosomes and seeds, though the elaiosomes nevertheless offer some protection for the seeds, as ordinary ants may remove the seeds before the harvesters find them.

Other creatures besides ants can be attracted to elaiosomes as a source of food.  In western North America a number of observers have reported seeing the common yellow jacket wasp Vespula vulgaris collecting seeds from Trillium ovatum and Vancouveria hexandra and carrying the seeds to their nests.  Apparently the wasps, like most ants, consume only the elaiosomes and discard the seeds, which makes them even better dispersers of seeds than ants because they travel greater distances.  Seed dispersal by wasps has come to be known as vespicochory.  One commercial grower told me that wasps are a serious pest for him as they remove Trillium seeds before he can get to them.

Myrmecochory in Erythronium (see more here)

At the time that little projections on the seeds of Erythronium japonicum made me wonder if some Erythronium were myrmecochorous, the idea was nothing but a wild guess.  There was absolutely nothing in the literature to suggest there was any connection between ants and Erythronium.  Applegate’s suggestion of there being two distinct groups of Erythronium was evidently largely intuitive.  To explore my idea I needed to see fresh seeds of Erythronium of the group that Applegate had considered distinct from the western species.  In my garden I had only species from western North America except for one clone of E. americanum, which flowered beautifully, but failed to set seed.  However, a garden shop had bulbs of E. japonicum, and a friend gave me a clone of E. americanum with which I could pollinate mine.  Now I should be able to have seeds of the group that might be myrmecochorous.  Already I was beginning to dream of the possibility that Applegate’s two groups might really be subgenera.  Everything  went according to plan and within a little over two years I had ripening capsules of both species.

Ant with spilled Erythronium seeds

Ant with spilled Erythronium japonica seeds

It was a very exciting morning when I found an E. japonicum capsule just beginning to dehisce.  I snatched it up, took a quick look at the seeds, which clearly had big, fat elaiosomes, and rushed to get my camera.  Within minutes I was at a spot where I knew ants would be scurrying about in the warm morning sun.  The very first seed I dropped near an ant confirmed my hopes, for the insect promptly seized it and set off towards its nest.  For the next hour or so I matched wits with the ants, as I tried to photograph them, and they tried to escape with the seeds, but then they began to lose interest.  The elaiosomes were losing their attractive power, and soon the ants paid no more attention to them than they would to pebbles.

At that point I remembered my E. americanum with its ripening seed capsules.  It did not take me long to reach the plant, and I immediately saw  that it was indeed my lucky day.  A capsule was just beginning to open and already several very tiny ants had found it and were trying to widen the opening to get at the seeds.  I quickly took up the capsule, shook off the tiny ants as being too difficult to photograph on that busy day, and headed back to my chosen ant arena.  The ants gave those seeds the same reception they had given those of E. japonicum.  Again I spent an hour or so taking photographs, and then again the seeds lost their attractive power.

I now felt reasonably sure I knew the major characteristic that separated the two groups identified by Applegate, but I was aware that a total of 10 species were known from Europe, Asia, and eastern North America, and I would need to ascertain they were all myrmecochorous.  That would not be easy to do, as I would need fresh seeds, and they can only be obtained by taking them from freshly dehisced capsules.  I live on the west coast of North America, so I would need to obtain and grow plants from what are for me distant parts of the world, and I would need to persuade the plants to flower and set seeds.  It took many years, for in most cases I could only obtain the plants by raising them from seed, which were often difficult to get, but eventually I had all the species except E. rostratum and E. propullans, both native to eastern North America.  As each species set seeds, I tested them as I had done with the seeds of the first two species, and with all of them I had equal success.  It was May of 2006 before the last of my species from far-away places, E. sibiricum, yielded its fresh seeds.

The two for which I could not obtain fresh seeds were not a serious problem.  Parks and Hardin (1963), who did a thorough study of the yellow-flowered Erythronium of eastern North America, concluded that E. americanum originated as a tetraploid hybrid of E. umbilicatum and E. rostratum, which tells us that E. rostratum must be closely related to the other two species and almost certainly would have elaiosomes on seeds that would be attractive to ants.  E., propullans could be a problem as it normally reproduces vegetatively, rarely produces seeds, and apparently does so only when pollinated from E. albidum.  Furthermore, it is an endangered species, so I could not obtain any seeds that it might produce.  However, a thorough study by Pleasants and Wendel (1989) concluded that E. propullans derived from E. albidum not over 9,000 years ago, so again I could be sure that the untested species must be closely related to a species that I had tested.

We cannot be absolutely sure that plants closely related to myrmecochorous species are also myrmecochorous, but my ultimate objective in studying these plants was not merely to determine what species are myrmecochorous.  I had become convinced that the two groups identified by Applegate correspond to two subgenera, and that one of those subgenera is distinguished from the other by being myrmecochorous.  A close relationship would certainly mean species belong together in the same subgenus.

The genus Erythronium divides naturally into two subgenera.

I am not the only person who has thought of dividing the genus Erythronium into two subgenera.  Shevock et al. (1990) described the two major groups as “perhaps corresponding to subgenera”.  However, to my knowledge, no published author has looked at the genus closely enough to describe the characteristics that identify the subgenera.  It was my good luck years ago that I spotted the projections on the E. japonicum seeds and that caused me to discover that the basic difference between the subgenera is that one is myrmecochorous and the other is not.

The Erythronium of western North America have their own method of dispersing their seeds.  Their firm-walled, cup-shaped seed capsules are held erect on tall, wiry stems, and when they are shaken by wind or a passing animal, the seeds are hurled out as from a catapult.  Not only the capsules, but also the seeds of western species are adapted to the catapult method of seed dispersal.  The seeds often must remain in the capsules for weeks or even months, waiting for the powerful shake that will hurl them forth.  That means the seeds must be very durable or they would be killed by exposure to the weather.  I once kept E. hendersonii seeds in a packet in my fridge for three years and got excellent germination from them when they were planted.

As one would expect, the capsules and seeds of myrmecochorous species are adapted to their method of seed-dispersal.  The capsules have leathery walls that after dehiscence curl away from the seeds, making them available to ants, and generally the mature capsules are either nodding or are prostrate on the ground.  The photos of a dehisced E. sibiricum capsule dropping its seeds compared with the erect, firm-walled dehisced capsules of E. revolutum illustrate the contrasting methods of seed dispersal of the two different subgenera.  There may be an exception to the general rule for the myrmecochorous subgenus.  I have not had the opportunity to study E. rostratum, but according to published information, its capsules are held erect at maturity.  That raises a question as to how the capsules dehisce to release the seeds.  One wonders if the species is vespicochorous; that is, if its capsules are adapted for the dispersal of the seeds by wasps.  As the range of the species coincides with that of harvester ants, there could be a considerable advantage to aerial transporting of the seeds.

Erythronium sibiricum from east of Novosibirsk

Erythronium sibiricum from east of Novosibirsk

There are other characteristics which distinguish the two subgenera.  The species of Europe, Asia, and eastern North America all produce only one flower per bulb, while those of western North America often have several.  Also the leaf mottling is different.  All the species of the first group have mottled leaves, though with both E. sibiricum and E. mesochoreum some individual plants and perhaps some populations have unmottled leaves.  The mottling in this group is formed of apparently random spots and blotches.  In the western group there are species with mottled leaves, and species with completely unmottled leaves, and two species with only a slight trace of mottling, evidently because they originated as hybrids between a mottled and an unmottled species.  (E. quinaultense and E. elegans are tetraploid species which apparently originated at some time in the past from the crossing of E. montanum and E. revolutum.)  The mottling in the western group has a somewhat symmetrical appearance, as it follows the veins in the leaf.

Having described the distinguishing characteristics of each of the two subgenera, it would seem sensible, for the sake of convenience, to name them.  The Myrmecochorous Subgenus, as I will call it, includes E. dens-canis, which was the first recognized Erythronium species, so if I could give it a scientific name, it automatically would be subgenus Erythronium.  It is almost as obvious that the subgenus of western North America — the subgenus that disperses its seeds by catapult action — would be named in honor of Elmer Ivan Applegate, who wrote a monumental monograph on the western species, published in 1935 and still of great value.  However, giving the subgenus a scientific name without providing a Latin diagnosis, and without having it published in a professional botanical journal or a book, which I am unable to do, would not fulfill the requirements of the International Code of Botanical Nomenclature, so I will be content with calling the western group Applegate’s Subgenus.

Each of the two subgenera has its own distinctive diagnostic features. 

  • Summarized from the above, the diagnostic characteristics of the Myrmecochorous Subgenus are as follows:
  • All species are native to Europe, Asia, or eastern North America.
  • The seeds have elaiosomes that adapt them for dispersal by ants.
  • The mature seed capsules generally either are prostrate on the ground or are held in a nodding position to release the seeds.  (E. rostratum is reported to be an exception.)
  • The mature seed capsules have leathery walls that rapidly curl away from the ripe seeds to make them available to ants.
  • The plants have no more than one flower per bulb.
  • Leaf mottling is generally present (except in some plants of E. sibiricum and E. mesochoreum).
  • The leaf mottling is in the form of apparently random spots and blotches.
Erythronium tuolummense

Erythronium tuolummense

The diagnostic characteristics of Applegate’s Subgenus are as follows:

All species are native to western North America.

  • The seeds lack elaiosomes and are adapted for exposure to the weather for weeks in an open seed capsule.
  • The mature seed capsules are held erect on strong, wiry stems and they have fairly firm walls that hold  the seeds as in a cup.
  • The ripe seeds are dispersed by being flung from the capsule when the stem is swayed by wind or a passing animal.
  • All species are capable of producing more than one flower per bulb.
  • Fewer than half the species (8) have strongly mottled leaves, and 2 species have merely a trace of  mottling; all other species (9) have plain green leaves.  (I recognize E. howellii and E. idahoense as species, though they are not recognized in the Flora of North America North of Mexico.)
  • The leaf mottling, when present, follows the veins of the leaf, and thus has a somewhat symmetrical appearance.

In the Myrmecochorous Subgenus the species often have quite different elaiosomes. 

In the photos of the seeds of E. dens-canis and E. japonicum one notices a striking difference in the shape of the elaiosomes of the two species.  E. dens-canis seeds have slender, curly elaiosomes that sometimes have the shape of a perfect corkscrew, while those of E. japonicum have a lumpy, round appearance that reminded me when I first saw them, right after returning from shopping, of a grocery bag stuffed with groceries.  The observation of those very different shapes is much on my mind.  I wonder if the difference is confirmation of the belief that the two are distinct species, rather than being two varieties of the same species, as botanists believed in the past, and as some may still believe.  That is, I am wondering if elaiosomes have a taxonomic significance in addition to that of helping to distinguish between the subgenera.  E. caucasicum and E. sibiricum both have curled elaiosomes, but they are much shorter than those of E. dens-canis, which may indicate that the two from western and central Asia are closely related, and have a more distant relationship to the European species.  It would seem that none of the three are closely related to E. japonicum, with its grocery-bag elaiosomes.

I only recently stumbled upon the idea of elaiosomes helping to distinguish between species, so the idea is mere speculation on my part until I have looked at the seeds of more plants of the above four species, and have had a careful look at the seeds of the species of eastern North America, which I could not do this year (2006) as various misfortunes prevented those species that are in my garden from producing ripe seeds.  However, it is my impression that the elaiosomes of the eastern North American species are very different from those of the Eurasian species.

The seeds of the two subgenera have different needs.

The seeds of Applegate’s Subgenus are durable and relatively easy to handle, though the ones from subalpine situations are a little more difficult.  With low-altitude species I prefer to sow the seeds in the fall in an outdoor seedbed with shade from deciduous shrubs and no drips from overhead trees.  A covering of very coarse sand provides protection from winter rains.  If the seeds are not planted too closely, the seedlings can be left in the seedbed until they flower.  High-altitude species may need a longer period of cool temperature than they would get outdoors.  They need to go into the fridge in the fall.  One can use the well-known moist paper towel method, though I prefer to use a 500-gram size of yogourt container, or similar container, with about 2 cm. of thoroughly moist sand in the bottom.  Everything should be clean, and the sand should be sterilized.  I sow the seeds on the sand and sprinkle on enough dry sand to not quite cover them.  (The dry sand instantly takes up moisture from the moist sand.)  A cover of thin plastic cut from a plastic bag will keep the moisture in but allow some passage of air.  The container goes into the fridge until the seeds germinate.  (On the bottom shelf of my fridge the temperature is about 4o C.)  Then I pot the germinated seeds in suitable soil, enclose the pot in a clear plastic bag which can be opened enough to provide necessary ventilation, and place the pot in a cool place with plenty of light but no direct sunlight.  Seeds that arrive from an exchange in midwinter are a problem.  As seeds of this subgenus are durable, I store them in the fridge until the following fall.  That may be better than having them germinate after warm weather has arrived, as Erythronium seedlings do not like warm conditions.

With the seeds of myrmecochorous Erythronium, a very different procedure is likely to prove most successful.  One must keep in mind the way ants “plant” the seeds and the time of year they “plant” them.  The ants leave the seeds on soil that is usually slightly moist in mid-summer.  That means the seeds get a fairly long period of warmth while moist before they receive the period of winter coolness which completes the germination process.  You can mimic that treatment.  I use sand in a yogourt container as I have described above, but instead of putting the container in the fridge, it goes into a moderately warm room.  About 22o C during the day and cooler at night seems about right.  About two months at this temperature seems necessary, but I have not done enough experimenting to be sure what period of time is ideal.  After the seeds have had their warm period, they go into the fridge and are treated the same as seeds of the other subgenus.  The above procedure worked very well for seeds of E. sibiricum that arrived in late October, were given two months of moist warmth, and went into the fridge about January 1st.  Almost all of those seeds germinated in February and early March.  I have not used this method for very long and seeds of the Myrmecochorous Subgenus are difficult to obtain, so I still have much to learn about them.  With fresh seeds of this subgenus from my own garden, I have given them a week to dry in a cool place after I collected them, and then I sowed them in an outdoor seedbed where they could have their necessary period of moist warmth.  That worked well.

While I discovered by watching ants how the dormancy of seeds of this subgenus is broken, others have learned the same thing simply by observing the seeds.  A 1985 paper by J.M. Baskin and C.C. Baskin, which I came across recently, gives very valuable information on the germination of E. albidum seeds.  Their study showed that E. albidum seeds are underdeveloped at the time they are dispersed in late May, and remain dormant under leaf litter during the summer months.  The embryos in the seeds gradually grow from early September through late January, with the most growth occurring in October and November, and with germination occurring in late winter and early spring.  Although they found that the embryos require low temperatures for growth, they also found that absolutely no growth takes place unless the seeds have first had the period of moist warmth during the summer months.  In brief, warm stratification followed by cold stratification is required to break the dormancy of E. albidum seeds.  That is exactly what I found with other species in the Myrmecochorous Subgenus.

Seeds of the Myrmecochorous Subgenus that arrive in midwinter can be an even worse problem than those of Applegate’s Subgenus as they may not be durable enough to survive until the following fall.  In 2006 a correspondent in Oregon reported that seeds of E. sibiricum and E. japonicum that she received in January were planted in pots that were placed outdoors, and they germinated well in late spring.  Perhaps those seeds were successful without any warm period, but it seems likely that either the February weather in Oregon was mild enough to give them something rather like a warm period, or perhaps more likely, the seeds were obtained from a well-informed dealer who had given the seeds their necessary period of warm stratification.

Erythronium seeds are unpredictable things because not only the different subgenera, but also different species and even different local ecotypes have evolved different needs.  One can only use one’s knowledge of the plants to try to improvise a treatment that will suit them.  As all Erythronium seedlings like cool conditions, in 2006 I tried to improvise a way to give late-germinating seedlings protection from the warmth of late spring and early summer.  I planted the late-arrivals in an outdoor seed bed with considerable shade.  Over them I placed a frame with a strong wire mesh supported about 10 cm above the soil, and on that I arranged narrow strips of wood to break the force of any sun that reached the bed.  Every evening without fail I sprayed the seedlings lightly with water.  Four groups of seedlings got this treatment, and although there were some losses, most stayed green long enough to make me hope they have formed strong bulbs.  (This is being written too early in spring for me to know my degree of success.)  Included in this experiment were a half-dozen germinated seeds of E. quinaultense I had collected while photographing the species in mid-May in its montane habitat north of Lake Quinault in Washington State.  I arrived home with those little green shoots, each with a seed at its tip, at the beginning of a spell of very warm weather, but they survived nicely.  Growing Erythronium from seed involves being very good at improvising.

continued in Part 2, here

References Cited

Anderson, A.N. and S.C. Morrison. 1998. Myrmecochory in Australia’s seasonal tropics: effects   of disturbance on distance dispersal. Aust. J. Ecol. 23:483-491.

Applegate, E.I. 1935. The genus Erythronium: a taxonomic and distributional study of the western North American species. Madroño 3:58-113.

Baskin, J.M. and C.C. Baskin. 1985. Seed germination ecophysiology of the woodland spring  geophyte Erythronium albidum. Bot. Gaz. 146:130-136.

Lanza, J., M.A. Schmitt, and A.B. Awad. 1992. Comparative chemistry of elaiosomes of three  species of Trillium. J. Chem. Ecol. 18:209-222.

Parks, C.R. and J.W. Hardin. 1963. Yellow Erythroniums of the eastern United States. Brittonia 15:245-259.

Pleasants, J.M. and J.F. Wendel. 1989. Genetic diversity in a clonal narrow endemic,            Erythronium propullans, and its widespread progenitor, Erythronium albidum. Am. J. Bot. 76:1136-1151.

Shevock, J.R., J.A. Bartel, and G.A. Allen. 1990. Distribution, ecology, and taxonomy of     Erythronium (Liliaceae) in the Sierra Nevada of California. Madroño 37:261-273.

About Art Guppy

Art spent over 70 years studying and writing about native plants of the Pacific Northwest from BC to California, especially the genus Erythronium and related plants. This site is a compilation of his work for the benefit of naturalists everywhere.
This entry was posted in Erythronium, Growing from Seed, Myrmecochory, Taxonomy and tagged , , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s