Lilium parryi, Lemon Lily, Fact Sheet

Fig. 1. Photographs of two different lemon lilies with an unusually large number of buds / flowers, in bud (left) and in bloom (right). See Life cycle of a Lemon Lily in Pictures.
Click on the pictures for larger versions.

This page presents some basic facts about Lilium parryi, in a brief format, to keep this page at a manageable size. A companion page elaborates on some of the information, and gives sources.

Table of Contents

Name Origin
Similar Species at San Jacinto Mountain
Population Characteristics
Plant Characteristics
Taxonomy and Phylogeny

Name Origin

Scientific Name origin: Described by Sereno Watson in the year 1877, published in Proc. Davenport Acad. of Nat. Sciences, Vol. ii, pp. 188-189, Davenport, Iowa 1880. Named for the collector, Charles Christopher Parry, a botanist with the Pacific Railway Survey.

Common Name origin: No common name was given in the original 1877 paper describing the species.

The first common name we have found in use for Lilium parryi was Parry's Lily, used first by John Lemmon in a letter to Sereno Watson on 6 August 1879: “In San Bernardino Co. I was fortunate in finding a secluded valley inhabited by Parry’s Lily and I brought away 300 bulbs. Can you help me to sell them?" (Wynne Brown email to Dave Stith on 30 March 2019). It was subsequently used by H.M. Hall in his 1902 Botanical Survey of San Jacinto Mountain, and in use through 1922 in Addisonia: colored illustrations and popular descriptions of plants on page 49.

The common name of Lemon Lily was first used by Theodore Payne in 1916 in California wild flowers, their culture and care on page 15. It was also used in 1917 by P. Beveredige Kennedy in Annotated list of the wild flowers of California on page 31. Both sources say the flower is "lemon-yellow", so the color of the flower almost surely accounts for their choice of lemon lily as the common name.

Jepson used lemon lily in his 1922 Flora of California and his 1925 Manual of the Flowering Plants of California, so that undoubtedly was the death knell of the name Parry's lily.

We had previously also speculated that the name lemon lily might have come from a corruption of, or play on, the name of J.G. Lemmon, who made a voucher one month prior to that of Parry. But given that Lemmon himself used the name "Parry's Lily" two years after the lily was named, that origin is ruled out.

The name "lemon lily" is also sometimes used for Hemerocallis lilioasphodelus, also known as "lemon daylily" and "yellow daylily". This is not just a theoretical confusion of the common names; it happened in one of the articles about the first lemon lily festival. See the differences between these two species.

Discovery locality: Boggy ground surrounding the house of J.G. and F.M. Ring at their mountain retreat near San Gorgonio Pass at 4000 feet elevation, where they grew potatoes.

Similar Species at San Jacinto Mountain

The most common species at San Jacinto Mountain that is sometimes mistaken as a lemon lily is the Humboldt lily, Lilium humboldtii. There is no problem in distinguishing these species when in bloom, since the flower of the Humboldt lily is yellow-orange with many spots, and pendent. However, vegetatively the plants are nearly identical, since both are true lilies. This is only a problem within Idyllwild proper, where they have been planted by some residents and sometimes have escaped into the creeks there. For some unknown reason, Humboldt lilies are not found naturally in the San Jacinto Mountains, even though they are quite widespread in southern California.

Sometimes Hooker's evening primrose, Oenothera elata, also abundant in Idyllwild and other San Jacinto Mountain locations, is mistaken for a lemon lily, since it is a tall plant with yellow flowers that grows in wet areas. However, it only has four petals, which form a flat, open flower, very different from our lemon lily with five petals in a trumpet form.

Population Characteristics

Geographic Range: Southern California: San Gabriel, San Bernardino, San Jacinto, Santa Rosa Mountains, and Palomar Mountain. Southeastern Arizona: Santa Rita, Huachuca, and Chiricahua Mountains (now presumed extirpated by fire). Mexico: extreme northern Sonora in the Sierra los Ajos. See large-scale map of the two disjunct populations.

Elevation Range: 4000 to 9550 feet. 4400 to 9550 feet at San Jacinto Mountain, with a single occurrence at 4400 feet (in the North Fork of the San Jacinto River at 5S09) and another at 4750 feet, and the rest above 5320 feet. The highest elevation plants are at the wet drainage on the Deer Springs Trail just south of the switchbacks up to Little Round Valley.

Habitat: Springs, seeps, wet meadows, and shady canyon bottoms along perennial streams, in areas protected from gophers.

Population: ~4000 plants in San Jacinto Mountains; ~3700 plants in the Huachuca Mountains; ~2,000(?) plants in San Bernardino Mountains; ~1300 plants in San Gabriel Mountains. Fewer than 100 plants in each of the other locations. (sources)

Conservation Status:
CNPS Rare Plant Rank 1B.2 = Rare or endangered in California and elsewhere; Fairly endangered in California.
California State Rank S3: Vulnerable.
Arizona Salvage Restricted.
Global Rank S3: Vulnerable.
USDA Forest Service Sensitive.
Not on the U.S. Fish and Wildlife Service Endangered Species List.

Major predators: Gophers eat the bulbs and deer eat the buds. Foliage is rarely consumed, probably by insects and possibly deer. Humans have nearly extirpated some populations by collecting the bulbs.

Threats to populations: Primarily human activities, such as bulb collectors, flower cutting, trampling, water diversion, grazing, introduction of exotic weeds, unnaturally frequent forest fires, development, logging and mining. If climate change reduces their available water, or causes more frequent fires, that will also have an impact.

Demographics: Although no one has ever done a longitudinal study to truly know the demographics of the population, we have enough data accumulated from the last eleven years that we now have a good idea of what constitutes a normal, healthy population or not.

A healthy population consists of about half plants too young to flower, and about half flowering plants. Of the flowering plants, about half produce just a single flower. The number of plants having increasingly-more flowers declines smoothly and rapidly; see Number of Flowers below. This decline is probably caused by mortality due to gophers, site changes, or bulb-splitting.

Because having seedlings present is highly correlated with having a number of young plants with a single bloom, it only requires one of those two observations to verify a healthy population.

The population in Candy's Creek was a healthy one from 2010 to 2019; its histogram of the number of plants with a given number of flowers, has remained the same in shape in surveys from 2010, 2011, 2012, 2013, and 2016. Baby plants making up about half the population have been observed in 2010 and 2019.

In contrast, the population in Skunk Cabbage Meadow, up to the Willow Creek Trail, is not a healthy population. Very few young plants have been observed there, and only 20% of the blooming plants have a single flower. Confirming this, the number of lilies along the Willow Creek Trail in this area, have declined with time, from about 40 plants in 2009 and 2010, to just 14 plants in 2020. Most of this decline occurred in 2012, when ten plants in just one spot were replaced with filled gopher holes. Gopher predation is probably quite normal here; the problem is that there has been little recruitment of new plants in the last decade.

The population in the burned portion of Willow Creek was a healthy population prior to the 2013 Mountain Fire, with about half the population consisting of young plants. After the Mountain Fire, the Creek was significantly affected by both erosion and choking amounts of sediment, and no young plants were observed in a 2018 survey of the upper part of the burned portion. It will be interesting to see how long it takes for this section to recover.

Alas, the population in Candy's Creek stopped being a healthy population in 2020 and 2021, when only 20% of the flowering plants had a single flower, instead of the expected ~50%. This is probably due to our changing climate, including the "megadrought" that began 20 years ago, the more erratic rain and snow, and global warming. Candy's Creek was completely dry in the 2021 survey, the first time in memory it has been completely dry in July. We have had a number of unusually-intense scouring storms at San Jacinto in the last decade, some of them also unusual in being warm rain storms, instead of the usual winter gentle snows. Three lemon lilies in Candy's Creek were lost sometime between 2011 and 2020 when erosion took away the stream bank in which they lived.

The demographics in the San Gabriel Mountains are similar. Mistretta and Parra-Szijj 1991 reported that they found "approximately 1275 total individuals" in their extensive 1990 survey there, with "677 (54%) of the total were individuals which were flowering or fruiting".

The demographics may be a bit different in Arizona, but the information is contradictory. One report was that in 1981, "60% of the observed plants in Arizona flowered and set fruit", which is nearly identical to our counts in Candy's Creek. However, detailed population counts in seven separate canyons showed that "seedlings and juveniles" made up 2/3 of the population and "fruits and mature" plants made up 1/3 of the population. (source)

Details to support our conclusions on the demographics at San Jacinto Mountain:

In our 2009 census of lemon lilies in Willow and Tahquitz Creeks, we estimated at the time that about half were young plants with one to five leaves, and half were flowering plants. But that was just an estimate, not based on actual counts.

We have since made two quantitative measurements that roughly confirm that distribution. First, in Candy's Creek, where we did surveys in five different years, in 2010 Dave counted 341 plants of all ages, and in a separate survey that same year, we counted 210 plants with a blooming stalk, making up 62% of the total population.

Second, in a repeat survey of two sections of Willow Creek, we found no young plants at all in the burn section, and a density of blooming plants about half the total number of plants pre-burn. This is consistent with young plants making up roughly half the population pre-burn. Furthermore, outside of the burn area, our estimate was that roughly half the plants were non-flowering. Most of those young plants were stalkless and had one or just a few leaves; "teenage" plants with a stem, but no inflorescence, were few in number.

Plant Characteristics

Life Form: Geophyte, from an underground bulb consisting of a series of scales; see a cultivated plant that had split into two bulbs.

Lifetime: Two plants have been documented to live for at least 18 years, and the very rare plants with 25 or more flowers may be as old as 50 years. Plants may split their bulb into two after ten years or so if they grown in good conditions.

We know of three bulbs that have split, with very different results. A potted plant grown by Kathy Bates-Lande only ended up with one robust stem after splitting, with a small number of flowers. The "grandfather" lily in Candy's Creek / Reeds Meadow produced 16 flowers in 2012, the year before it split, and then produced two equally robust stems after splitting, with one stalk with six flowers and the other with three flowers, in 2019. It was completely absent in 2020, with no evidence of gopherization. The "Big Momma" lily in Idyllwild produced 26 flowers in the year before it split, then had one stem with 17 flowers and another with 15 flowers the next year. Unfortunately, "Big Momma" was no longer seen after that year.

A repeat survey in a post-burn survey along Willow Creek, where seedlings were not present for at least five years, indicate that the flowering plants there are at least seven years old, with little evidence of mortality during that period.

The report that lemon lily plants live only a few years after flowering comes from cultivated plants, which almost surely are not growing in optimum conditions.

Known Mortality Events: gopher predation (two events; a plant just below Little Tahquitz Meadow taken out within a few months after bloom in Summer 2020, replaced by gopher mounds; at least eight plants, and maybe more, taken out in the bracken meadow, replaced by gopher holes); bulb-splitting ending in death (two events; see Lifetime just above), creek bank erosion (three plants in one location; see Natural Deaths of Lemon Lilies: Erosion); smothering by creek deposits after the 2013 Mountain Fire (many now missing from repeated surveys).

Seed Germination: Seeds require a cold period of approximately 45-50° F for about two months in order to produce the first true leaf above ground. See development of a seedling.

Time from Seed Germination to Flowering Plant: three to five years depending on weather conditions, proximity to a water source, etc. Some lilies may never bloom in their lifetime if conditions are not favorable. The munchkin lemon lily, growing in an unreliably-wet area, has only produced a flower stalk three times in sixteen years of observation from 2003 to 2018, in 2003, 2011 and 2012.

Roots: All true lilies have contractile roots (thick and concentrically wrinkled) and thinner, fibrous roots. The contractile roots move the seedling bulb down in the soil, aided by the displacement of soil done by the thick roots, to reach its optimum depth. See a bulb that had split into two bulbs with all of its roots, and the two bulbs after being separated showing the roots more clearly. A young plant from a bulb scale has a very long contractile root, with only the uppermost part being able to contract, and the much longer lowermost part serving to anchor the root to pull down the bulb.

Stem: Seedling plants have no stem, just a single leaf. Second year plants usually produce multiple leaves from below ground. Older plants usually produce a single stem from the center of the bulb with alternate cauline leaves. Stem height of blooming plants ranges from two to nine feet.

Leaves: Young plants and most non-blooming plants have alternate cauline leaves. A few non-blooming plants end their growth with a whorl of leaves. Blooming plants almost always additionally produce one or more whorls of leaves, with up to 18 leaves in a whorl, above their alternate leaves. Typically, a blooming plant has two to four sets of whorled leaves, usually three, with seven to eight alternate leaves, although small plants may have only a pair of opposite leaves below the flower, and we have seen a single extremely short plant with only alternate leaves. Above the whorl(s) of leaves, the flowers have leaf-like bracts that are alternate, opposite, or whorled; see the Inflorescence section below.

The maximum number of leaves and buds in a given year appears to be determined by the size of the bulb, and by conditions in the previous year. As for all lilies, and many other flowering plants with bulbs, the leaves and buds for next year's plant are produced in embryonic form in the bulb by the plant in the previous year. If the plant was damaged in the previous year, or if the plant didn't reach its full potential that year due to drought or other growing conditions, the bloom will be poor in the following year.

If conditions are good in a given year, the plant will produce that maximum number of leaves and buds that was determined in the previous year. If conditions are not good, the plant may not produce any buds, nor the maximum number of leaves it might have had. Even in a good year, if something happens to the leaves or buds (e.g., herbivory or frosts), they can not grow more to replace the damaged parts.

Inflorescence: As discussed immediately above, the entire inflorescence was created in embryonic form by the plant during the previous year. This "born with all the parts it will ever have" is apparent when the plant starts to form its inflorescence. See the entire set of flower buds and bracts in a developing inflorescence, which eight days earlier were hidden inside the developing leaves.

Plants large enough to bloom typically produce just a single terminal flower in their first years above either a single leaf or bract, or a pair of opposite leaves or bracts. As plants grow more robust, the single flower emerges from a whorl of a number of leaves; see the rightmost plant here. Plants with more than one flower produce one to two leaf-life bracts under each flower in the inflorescence. Plants with two to three flowers usually have their flowers umbellate (coming from the same node on the stem; see the leftmost two plants in this pix). Plants with four flowers usually have two to three flowers coming from the lowest node, and one to two flowers coming from the next node. Plants with a very large number of flowers can have alternate flowers, or many flowers per node, usually in a pyramidal shape, with the lower flowers having longer pedicels (flower stalks).

The lowest flowers on the stem open first.

Progression of Flowers and Fruit as they develop: Buds emerge from the top of the stem erect, and become pendent as they enlarge and the pedicels elongate. Flowers are spreading (horizontal) to somewhat nodding at bloom. The fruit becomes erect.

Variation in the number of flowering plants with year: We and others have surveyed the same area of Candy's Creek in five different years. The number of flowering plants varies strongly with conditions in a given year. The observed numbers were 214, 170, 38, 127, and 178 in the years 2010, 2011, 2012, 2013 and 2016 (a post-fire year). In 2011, 26 plants had buds aborted due to a late frost, so the total would have been 196 that year without the frost.

The years 2012 and 2013 stand out with their low numbers. Both of those years had well-below average rainfall. Although 2013 had a much lower rainfall than 2012, 2012 had by far the fewest plants that bloomed. This is consistent with the potential number of blooms in a given year being determined by conditions in the previous year. From rainfall alone, 2014 should have had the fewest blooming plants. That may well have been the case, but we weren't allowed to do our survey due to the Forest closure after the Mountain Fire. Perhaps the frost in 2011 affected the development of the embryonic buds for the bloom in 2012.

Flowering time: Lemon lilies have been observed to be in bloom in at least one year in the following months. In Arizona, May and June, with June being peak bloom. In southern California, June and July at elevations of 4000 to 6000 feet; and July to August at elevations of 7000 to 9000 feet.

However, the actual bloom duration at a given elevation in a given year is typically something like two weeks, much shorter than this range of two months from all years combined, since weather patterns in a given year can cause them all to bloom early in that two month range, or for all to bloom late. For example, at 8000 feet elevation in the San Jacinto Mountains in 2019, a colder and wetter year than normal, lemon lilies were observed to be in bloom from 22 July to 1 August, with peak bloom around 26 July. Lemon lilies were observed to have only buds on 15 July, and from observations, the last flower on plants with a large number of flowers probably finished blooming on 2 August.

Anatomy of a Flower: See the parts labeled by Dave on a photograph by Bob Herrmann.

Number of Flowers: Among blooming plants, plants with a single flower are most common; just under half of all flowering plants have only a single flower. The number of plants having increasingly-more flowers declines smoothly and rapidly, with 90% of all flowering plants having one to four flowers, and 98% of all flowering plants having one to eight flowers. At that point the smooth decline stops, and there is almost a uniform distribution of plants having nine to 16 flowers, with only a very small number of plants in that distribution. An extremely-small number of plants produce anomalously-more flowers, up to 51 flowers for the Champion Lemon Lily, and ~25 flowers for the runner-ups.

See a histogram of the number of plants with a given number of flowers, compiled from the sum of five years of surveys (from 2010, 2011, 2012, 2013 and 2016) in Candy's Creek. Each of the five years shows the same pattern, despite a variation in the number of flowering plants from a low of 38 plants in 2012 to a high of 214 plants in 2010.

This distribution in the number of flowers probably represents the influence of several factors, such as habitat quality (amount of sunshine, water, and nutrients), and the turnover of the population (the percent of plants that die at a given age from predation, disease, or change in habitat quality). Older plants in good habitats that are also protected from gophers will produce large numbers of flowers year after year, at least until their bulb splits which reduces the number of flowers on the two new stalks. Plants in poor habitats may consistently produce only a few blooms, no matter the age of the plant. Plants that are not protected from gophers will have a high mortality rate and be constantly replaced by new plants with few flowers.

The average number of flowers per blooming plant is 2.3, since plants with more blooms contribute more flowers to the average than plants having a single flower.

Time Flowers Open: Buds generally open just before dusk, then the flowers stay open until they fall apart. Deb Nelson photographed a bud that began opening within 20 minutes after 2:35 p.m. on July 3, 2016.

Individual Flower Lifetime: It depends on temperatures. Dave observed blooms that last four to five days; Kathy Bates-Lande observed blooms lasting about one week. Kathy's flowers were all pollinated from the time they opened, twice a day, so pollination does not seem to decrease the lifetime. Note that essentially 100% of lemon lily flowers get pollinated.

Fragrance of flowers: a strong very-pleasant fragrance that is hard to describe, but which is not that of a lemon. The flowers can sometimes be smelled from standing next to them in their first few days. The fragrance fades in older flowers.

Pollinator: Sphinx moth, Hyles lineata; possibly also other insects and hummingbirds. It is one of only two Lilium species that independently evolved to be pollinated by hawkmoths, the other being L. washingtonianum. Grant (1983) listed just 105 species in 16 families that have hawkmoth flowers (ones with a nectar tube of generally 2 to 7 cm, in some cases up to 9 to 20 cm) in North America north of Mexico. See more info.

Number of Flowers Producing Fruit: Before the year 2020, our observations from a span of a decade, were that close to 100% of flowers produce fruit, in areas with a number of flowering plants. This was true even though the flowers are said to be strictly outcrossing, and must be pollinated from another plant. The last flower to bloom in an area is often not pollinated, since there are no other flowers around to pollinate it. Kathy Bates-Lande reports that occasionally the last few flowers do not develop a stigma, but do contain anthers with pollen. A lily grown in isolation will produce no fruit, and areas with a small number of plants producing few flowers might produce few fruit.

Mistretta and Parra-Szijj reported that the fruit set was nearly 100% in the San Gabriel Mountains in 1991, and Skinner reported this was typical in his 1988 Ph.D. thesis on lilies.

However, in 2020, we observed that the fruit set was markedly lower, just ~half, in five separate locations. There was an interesting pattern in the data; plants with just one or two flowers set just 14% of them with fruit; plants with three flowers set 44% with fruit; plants with four to eight flowers set 57% with fruit; and plants with ten to 13 flowers set 68% with fruit. For lilies with multiple flowers, it was still true that the first and last flowers were more likely to be unpollinated. But this wasn't strictly the case; some middle flowers did not set fruit. Our best speculation is that this low fruit set was due to fewer blooms this year, along with fewer pollinator visits.

Predation of Fruit: Most fruit develop untouched. Deer chomp the buds / flowers / fruit from 10 to 20% of the plants in surveys we've done in Candy's Creek. A much-smaller number of fruit (much less than 1%) are damaged by either insects or birds. Since 2009, we've observed such damaged fruit only three times: Four damaged fruit at 9600 feet elevation in the Lemon Lily Drainage just below Little Round Valley on 19 September 2010; three damaged fruit at ~8200 feet elevation in Willow Creek above the Willow Creek Trail crossing on 8 August 2018; and a single damaged fruit at 7940 feet just below Little Tahquitz Meadow on 16 October 2020 (see also iNat observation). Similar fruit damage was observed in Michigan on a sister Lilium species.

Number of Seeds per Fruit: 140 to 300.

Dispersal of Seeds: Many seeds simply fall to the ground as the capsule splits open or the stalk falls over, from our observations of baby lilies at the base of parent plants. Gusts of wind can also pick up seeds in the capsule and disperse them to somewhat farther distances; see photograph of seeds dispersed by Dave blowing on a capsule at the bottom of Fig. 1. If the seeds fall on water, they can be dispersed to larger distances either by being blown over the water surface by wind, or carried away in the current, since the seeds are hydrophobic and float on the water surface.

Viability of seeds: At least five years when stored clean and dry at ambient room temperature (from Dave's germination tests). We do not know what the viability of seeds in the wild is; it might be only a single year.

Response to fire: Flood deposits and erosion following the 2013 Mountain Fire along Willow Creek wiped out all the young lemon lilies for at least the first five years after the fire, leaving only the older, well-established plants, from a 2009 survey before the fire and a 2018 survey after the fire. See plot of the number of plants at each GPS point vs. elevation in Willow Creek. Above the burn area babies and young plants were present, reflected in the increased numbers of plants at many of the GPS points; in the burn area only adults were present, the vast majority of which were flowering.

There are reports of populations in Arizona that were thought to be extirpated by floods, but which were rediscovered some years later. There was one report that a flood eradicated a patch of Equisetum allowing some lemon lily seedlings to become established.

Taxonomy and Phylogeny

Number of Lilium species in the world: ~100.

Number of Lilium species in North America north of Mexico: 22.

Number of Lilium species in California: 12 species; 18 total taxa including subspecies.

Number of Lilium species in Arizona: just one, L. parryi.

Origin of Lilium species: eastern Asia, with about 60 extant species remaining there.

Sister Genera Sharing a Common Ancestor: Fritillaria.

Evolutionary Group in North America: Six species of moist or wet places with rhizomatous bulbs that frequently branch, which apparently originated from within the dry-ground group that consists of six species. (The bulbs of L. parryi do not branch except rarely for very old bulbs. Their rhizome is inside the bulb scales and minutely moves the flowering stem of the bulb each year.)

Combination of traits that distinguishes L. parryi from other North American species: Flowers nodding to spreading; sepals and petals bright yellow, not distinctly clawed.

Closest relative: probably L. pardalinum, despite the very different appearance of its flowers (with pendent orange flowers with reflexed sepals and petals). L. pardalinum mostly lives to the north of the Transverse Range, although there are plants in the Laguna Mountains and perhaps at Palomar Mountain. L. parryi essentially replaces L. pardalinum in its range.

Relative with the most-similar-in-form flowers: L. washingtonianum, with white flowers that are remarkably similar in form and function and which are also moth-pollinated. This resemblance is due to evolutionary convergence. In California, L. washingtonianum lives north of Fresno in the Sierra Nevada and Klamath Ranges.

Bogus variety: var. kessleri, claimed to be distinguished by its wider leaves in the San Gabriel Mountains, had wider leaves simply due to heavily-shaded habitats. We've seen similarly-wide leaves in heavily-shaded habitats in the San Jacinto Mountains.

Go to:

Copyright © 2018-2022 by Dave Stith and Tom Chester.
Commercial rights reserved. Permission is granted to reproduce any or all of this page for individual or non-profit institutional internal use as long as credit is given to us at this source:
Comments and feedback: Tom Chester
Updated 26 July 2022.