Lilium parryi

The jointed bulb-scales and rhizomatous bulb distinguish this species from L. COLUMBIANUM and L. HUMBOLDTII (Berg 1989). The leaf arrangement aids in distinguishing this species from another California lily, LILIUM PARDALINUM. Synge (1980) reports that leaves higher on the stem tend to occur in an alternate arrangement and leaves at the basal portion of the stem are whorled on LILIUM PARRYI. This leaf arrangement was not consistently seen in either Arizona or California (Berg 1989, Newman 1989a&b, Wood 1989), thus it is a poor descriptive aid for distinguishing this species.

LILIUM PARRYI occurs in wet soils of mountainous terrain, generally in forested areas, between 5,000 feet and 9,000 feet elevation. Lemon Lily is usually found growing along the shady edge of a stream but also occurs on seep water-replenished walls and, in California, in open boggy meadows (Neel 1989, Warren et al. 1989, Wood 1989). This species is found in mixed conifer forest, broad-leaved deciduous riparian forest and pine-oak woodland plant communities (Babb 1986).

Lemon Lily is the only true lily growing in Arizona. Two of the eight populations known to have existed in Arizona were in the Santa Rita Mountains: the Rosemont plants were extirpated due to mining in the early 1900s (Johnson 1988), and only one plant remained in Madera Canyon after the 1983 flood. This plant has not been seen for several years (Kaiser 1989). Several newly located populations were found during the summer of 1989: one population in both Florida Canyon and Huachuca Canyon, and a single plant in Saw Mill Canyon (Riggs 1989c, Warren 1989). The Huachuca Mountains are home to the remaining five populations which occur in Miller Canyon, Bear Canyon, Carr Canyon, and Ramsey Canyon (Johnson 1988).

Information on population size in Arizona is sparse. The Bear Canyon population data appears to be missing in Toolin's 1982 total plant count of 210 to 220 (Toolin 1982). Over 500 plants were found in both 1988 and 1989 (Johnson 1988, Newman 1989a&b, Riggs 1989c, Wood 1989).

The Arizona populations all occur in riparian environments at elevations greater than 5500 feet in shady canyon bottoms. The plants grow along the edge of perennial streams in pine-oak, coniferous forest and broad-leaved deciduous plant communities. The upper Ramsey Canyon population grows next to a stream on a steep hill fed by seep water. A major threat to Lemon Lily arises from its dependency on water. Fluctuations in the water level may lead to a decrease in the soil's water potential and to erosion of the streambed (Toolin 1982, Babb 1986, Johnson 1988, Wood 1989).

Most, but not all plants, are imbedded in two to six centimeters of leaf-litter and usually in organic soils. An exception is the lower Ramsey Canyon population which occurs, not in organic soils, but in mineral soils (Heitlinger 1989). Saturated soil conditions exist for most of the year (Toolin 1982, Wood 1989). The soils in the Petran Montane Conifer Forest communities tend to be slightly acidic, but due to the limestone substrate some soils where LILIUM PARRYI grow are alkaline with a pH of 7.8 (Wood 1989). Miller Canyon plants grow in soil with a pH of 6.8 (Deecken 1989). The associate species in Arizona include PINUS PONDEROSA, ABIES CONCOLOR, ACER GRANDIDENTATUM, QUERCUS HYPOLEUCOIDES, PSEUDOTSUGA MENZIESII, HABENARIA LIMOSA, RUDBECKIA LACINIATA, AQUILEGIA CHRYSANTHA, FRAXINUS sp., MIMULUS GUTTATUS, PLATANUS WRIGHTII, ACONITUM sp. and EQUISETUM sp. (Toolin 1982, ANHP 1989, Riggs 1989a).

Approximately 15 populations, comprised of 2,000 to 3,000 individual plants, grow in California. Lemon Lily populations have been found in the following areas: Big Cienega Spring in the San Gabriel Mountains, Los Angeles County; Fuller Ridge Trail, North Fork of Stone Creek, and Dark Canyon in the San Jacinto Mountains, Riverside County; San Gorgonio Pass "Ring Brothers Potato Ranch" - type locality (Parry 1878), Dry Lake, South Fork Meadows, Horse Meadow, Big Bear Valley, Fish Creek Meadows, North Fork Meadows, Upper Whitewater River and recently at Gordon Springs (Foster and Neel 1988) in the San Bernardino Mountains, San Bernardino County; in the Palomar Mountains and Volcan Mountains, San Diego County, where they are rare (Toolin 1982, CNPS 1988, Dice 1989, Hamilton 1989, Neel 1989, Sanders 1989). Cattle grazing recently decimated a population at Cedar Spring in the Santa Rosa Mountains in San Diego county (Sanders 1989, Warren et al. 1989).

California Lemon Lily populations grow at elevations between 5000 feet and 9000 feet, in open boggy meadows, by streamsides, and on seep water-replenished walls. An exception in the elevation range of this species is one of the populations in the San Jacinto Mountains which occurs at 4400 feet and grows in sandy soil along the North Fork San Jacinto River (Berg 1989, Hamilton 1989). The other San Jacinto Mountain populations grow in open, sunny, wet meadows (Berg 1989, Hamilton 1989). The San Bernardino Mountain populations grow in wet meadows, by seeps and along streams, generally in Yellow Pine and Lodge Pole Pine subalpine forests in both open and densely vegetated areas (Berg 1989, Neel 1989, Sanders 1989). The plants growing at the type locality occurred in "boggy ground" (Parry 1878). According to Foster and Neel (1988), the most recently located population at Gordon Springs occurs in an unlikely site. The plants grow on a 20o-45o north-facing slope on limestone soil under filtered sunlight in saturated soil near a desert spring surrounded by a Canyon Live Oak stand. LILIUM PARRYI grows in rich boggy meadows in San Diego County. The San Gabriel Mountain population grows along a creek edge (Berg 1989).

In California, Lemon Lily plants tend to grow in the Sierran Montane Conifer Forest biotic community on sandy granitic soils with a pH of approximately 6.0 (Toolin 1982, Hamilton 1989). The dominant associate species in California are PINUS PONDEROSA, ABIES CONCOLOR, QUERCUS spp., ASTER spp., VERTRUM CALIFORNICUM, PTERIDIUM AQUILINUM, HABENARIA DILATATA, AQUILEGIA FORMOSA, JUNCUS sp., CAREX sp. and EPICACTUS GIGANTEA (Toolin 1982, Skinner 1988, Neel 1989, Riggs 1989a).

Populations and Density: Throughout this report population size and seedling numbers are reported for single rooted stems; however it should be kept in mind that asexual reproduction via the rhizome is likely. Deecken (1989) feels that vegetative propagation is more significant than sexual reproduction in the Miller Canyon population. Newman (1989a&b) believes that consideration of the rooted stems as separate genetic units is reasonable because of the clumped distribution of the single-leafed "seedlings," this pattern resembles seed scattering far more than rhizomatously derived stems.

Population size in California and Arizona varies considerably from several plants to several hundred plants. The average density in local patches is three plants per square meter (Toolin 1982). Ascertaining the number of individual plants is difficult due to the possibility of rhizome-produced stems. Table 1 illustrates the demographic patterns of the Arizona populations. The data was collected during the summer months between May and July; the month of collection is not known for all of the locations. Comparable information is unavailable for California.

Table 1. Arizona LILIUM PARRYI population data.

a) Ramsey Canyon DATE JUVENILES SEEDLINGS FRUITS & MATURE 1985 105 13 51 1987 52 26 20 1988-May 28 0 7 1989-May 41 22 2 1990-June 32 127 2 (Deecken 1989, Warren et al. 1989, Wood 1989, Gori et al. 1991)

b) Miller Canyon DATE JUVENILES SEEDLINGS FRUITS & MATURE 1986-June 200 800 ? 1988 403 ? 270 1989-May 316 376 46 1990-June 1036 1106 78 (Babb 1986, Warren et al. 1989, Gori et al. 1991)

c) Bear Canyon DATE JUVENILES SEEDLINGS FRUITS & MATURE 1988-July 60 ? ? 1989-May 70 149 7 1990-June 115 114 33 (Warren et al. 1989, Gori et al. 1991)

d) Carr Canyon DATE JUVENILES SEEDLINGS FRUITS 1987-July 73 ? 1988 57 ? 38 1989-June 117 48 64 1990-Junw 138 71 53 (Riggs 1989c, Warren et al. 1989, Gori et al. 1991)

e) Huachuca Canyon DATE JUVENILES SEEDLINGS & MATURE 1989-June 187 796 (Gori et al. 1990)

f) Florida Canyon DATE JUVENILES SEEDLINGS FRUITS & MATURE 1989-June 18 1 14 1990-June 36 21 24 (Gori et al. 1990, Gori et al. 1991)

g) Madera Canyon DATE JUVENILES SEEDLINGS FRUITS & MATURE 1990-June 16 15 10 (Gori et al. 1991)

The Arizona populations are situated in locations where changes in the stream level could influence the survivability of the plants. Severe floods following large forest fires appear to have detrimentally affected the populations of Carr Canyon in 1977 and Ramsey Canyon in 1983. The Carr Canyon population was thought to be decimated by the 1977 flood but was rediscovered in 1979. Boulders protected a plant cluster during the high waters of the 1983 storms (Johnson 1988, Collazo 1989).

The Ramsey Canyon population size has steadily decreased since the downcutting of the stream caused by the 1983 floods. Nine of the 118 plants labeled in 1985 survived through 1988. A total of 28 plants and seven fruits were present in 1988, whereas the 1985 population possessed 51 fruits (Warren et al. 1989). Only two flowers were observed in 1989 and these were hand-pollinated, producing 2 fruits; two flowers were again observed in 1990 and both set fruit with natural pollination (Deecken 1989;Gori et al. 1991). The decrease in number of plants and quantity of fruit production over the five years of monitoring is alarming.

The Miller Canyon and Bear Canyon populations have increased in size since 1988. Fruit production has also increased over this period at Bear Canyon, however, at Miller Canyon, the number of fruits was greatest in 1988 because of the low incidence of herbivory on flowers and stems that year. The herbivores are most likely bear and deer.

In 1989, a new population was found growing on a cliff, 60 feet above the stream, in Huachuca Canyon. Seventy-nine stems of the 187 plants were broken, possibly by bears; adjacent to the plants was a bear trail. Almost 10% of the plants were flowering (Riggs 1989c; Gori et al. 1990).

Two other new locations were found during the summer of 1989. A single, three foot tall plant, with five flowers, was found growing near a spring in Sawmill Canyon by Fort Huachuca wildlife biologist Clark Derdeyn (Riggs 1989c). Bob Perrill discovered a population in Florida Canyon in the Santa Rita Mountains (Warren 1989, Gori et al. 1990). In 1990, Sue Rutman rediscovered a population in Madera Canyon that was thought to be destroyed by a major flood in winter 1983.

Habitat Model Hypotheses: Since data preceding major floods is not available, we can only speculate on how populations may vary with cycles of wet and dry years in southeastern Arizona. It appears, however, that at least the Ramsey Canyon population was relatively strong shortly after the last major floods in 1983, and declined since. Collazo (1989) believes flooding may favor Lemon Lily by clearing away other vegetation, especially EQUISETUM, and thereby favoring Lemon Lily germination. Others speculate that downcutting in Ramsey creek by the 1983 floods lowered the water table beneath the lilies, causing declines in the population of adult plants. Such downcutting is not seen in Miller Canyon, where the population would appear to be more stable. This suggests a model similar to one developed by Bowles and Apfelbaum (1989). In this model, the plant benefits most by an intermediate amount of flooding, but declines as floods are reduced to near zero or increase to a high intensity or frequency.

Three alternative models have been constructed to help analyze the connection between fires, floods and Lemon Lily population size in Ramsey Canyon: inherent stability model, inherent instability model and altered flood frequency model (Heitlinger et al. 1989). The inherent stability model suggests that this system was essentially stable prior to human intervention in the form of fire suppression. Excluding fires from the watershed resulted in unnatural fuel buildups, and catastrophic results when wild fires occurred. Fires which would have been confined to the forest floor under natural, low fuel conditions, instead spread through the canopy of ponderosa pine. The subsequent heavy rains in 1983 caused more extreme flooding and erosion than they would have if the fire damage had not occurred. This model is supported by evidence that Miller Canyon was exposed to the same 1983 rainfall but did not suffer stream downcutting. However, the differing flood intensities may possibly be related to the physical characteristics of the Miller and Ramsey watersheds.

The inherent instability model suggests that high intensity flooding, and consequent wide fluctuations in Lemon Lily populations, are a normal phenomenon. In this interpretation, Lemon Lily numbers would regularly be reduced to the numbers observed currently in Ramsey Canyon. Frequent reduction in population size to a few individuals, would suggest that within population genetic variation should be low. Intervention would not be advised unless other effects of human practices have resulted in an inability of the population to recover after flooding (Heitlinger et al. 1989).

The third model is similar to the inherent instability model. The altered flood frequency scheme suggests that due to human intervention, such as fire suppression, the interval between naturally occurring floods has been reduced. There is less chance for seedling establishment when floods come at a greater frequency. In this case, intervention is necessary to reduce stresses due to the shortened intervals between floods, although periodic, high intensity floods, are anticipated (Heitlinger et al. 1989).

Unfortunately there is no available information on the relationship between flood cycles and LILIUM PARRYI demography in California. However, flooding is considered a potential threat to plants growing on sandy soil in California (Hamilton 1989).

Vegetative Propagation: Members of the genus LILIUM reproduce from seed, bulb and rhizome (Craig 1928). The nature and frequency of vegetative and sexual reproduction in the wild is not known for L. PARRYI. Since the mid-1800s lily breeders have been experimenting with Oriental, European and American lilies, determining the optimum hybridization and propagation techniques (Craig 1928, Grove 1935, Synge 1980).

Much of our knowledge of Lemon Lily's reproduction comes from cultivated plants. If current techniques had been available when this species was in demand Lemon Lily would, most likely, have been propagated using tissue culture at Oregon Bulb Farms (McCray 1989). Tissue culture allows for the most rapid production of genetically 'superior' ornamental plants (McCray 1989). All forms of vegetative propagation of cultivated lilies, including LILIUM PARRYI, are rapid and easy, expediting the process of obtaining flowering-sized plants. Scales removed from a bulb are laid flat, with the concave side facing up, in a peat/sand/loam soil mix, and within a year Lemon Lily bulbils are formed (Kline 1989). Specific information on rhizome propagation of LILIUM PARRYI is unavailable. A few lily species that normally do not produce stem bulbils (such as L. PARRYI) can artificially be forced to do so by removing the top of the plant during early floral development (Synge 1980). Possibly this occurs naturally in response to insect damage to the inflorescence or stem.

Cultivated Plant Sexual Reproduction: Lilies are self-sterile, and according to lily breeders, require large genetic variability for healthy, strong plants (Kline 1989, McCray 1989). Ripe lily fruit contain approximately 200 seeds (Synge 1980). A total of 135 seeds and 160 seeds were found in the flowers of two different LILIUM PARRYI plants growing in Ramsey Canyon (Wood 1989). Seeds keep for one season at room temperature, but cool temperatures prolong viability. Refrigerator and freezer storage increases the shelf life of seeds from several years to 20 years, respectively (Kline 1989). Breeders find a 90% germination rate after the undeveloped seeds (approximately 30%) are removed (Kline 1989). Currently several botanical gardens are having difficulty germinating seeds of this species (Riggs 1989b). Cool temperatures of 50 F to 60 F are essential in order to obtain maximum germination (McCray 1989).

Cultivation Technique: LILIUM PARRYI's hypogeal germination process requires cool temperatures for the production of roots and a bulb during the first season of growth (McCray 1989). During the following spring one true leaf develops, and three to five years after germination a flower stalk is produced (McCray 1989). A healthy, horticulturally sound, seed bearing plant will live for several years (McCray 1989). In order to prevent bulb rot, watering is terminated from the time the bud appears through floral senescence (Kline 1989). Every spring, reproductively mature plants are treated with a high phosphorous, low nitrogen containing fertilizer (Kline 1989). The plants are easy to grow if they are kept cool with saturated soil conditions below the bulb, at the root zone. High temperatures cause the plants to wilt and seed production to terminate (Grove 1935, Synge 1980, Kline 1989, McCray 1989). Cultivated plants perform better when grown in large groups rather than in rows, possibly because they shade each other (Graaff 1970).

Hybridization: Lily plants hybridize readily and thus breeders keep different species far apart. When cross pollinating, pollen should be collected from freshly dehisced anthers (Synge 1980). LILIUM PARRYI has been crossed with many other species, including eastern and New Zealand lily species, to produce assorted hybrids such as 'San Gabriel' (Graaff 1970, Synge 1980, Kline 1989, McCray 1989). Several lily breeders and taxonomists believe that 'new' species in the wild are actually hybrids. L. HUMBOLDTII readily crosses with L. PARRYI in cultivation and possibly in the wild (Hamilton 1989, McCray 1989). To prevent genetic swamping of the species this hybridization potential should be considered when reintroduction plans are prepared.

Pollination: Because Lemon Lily is self-sterile a pollinator is required (Kline 1989, McCray 1989). The primary and secondary pollinators of LILIUM PARRYI in the San Bernardino Mountains are two species of hawkmoth, HYLES LINEATA and SPHINX PERELEGANS, which frequent the plants at dusk. Several species of hummingbirds, noctuid moths, and many hymenoptera species were seen visiting the flowers throughout the day (Skinner 1988). The pollinators of the Arizona populations may be the same as those in California, but this has not been determined.

Seedling Establishment: Plants in Arizona flower and set fruit from June through July. California floral production takes place a few weeks later (Toolin 1982). In 1981, 60% of the observed plants in Arizona flowered and set fruits, but according to Toolin (1982) only 5% of the seeds germinated. Weakening of the peduncle by insect predation possibly contributes to germination failure (Toolin 1982). Since Lemon Lily grows along streams, the most likely mode of dispersal is through the water. This unreliable method of locating a good stable site for seedling establishment could account for the extremely low recruitment success found in the wild by Toolin (1982).

As a seedling becomes established, it remains dependent on a relatively constant stream level. The plants often wilt and die when the water level drops, but the fragile soil along the water's edge is swept away when the water level rises. Established seedlings have a low survival rate in Ramsey Canyon (Wood 1989). In some instances high water levels may be beneficial to seedlings by clearing the area from competition, allowing the plants an early start in their development. Tom Collazo (1989) thinks the flood of 1983 in Ramsey Canyon was beneficial, at least in the short-term, because it cleared the EQUISETUM sp. allowing many new seedlings to become established.

Insect Damage: A true bug, possibly a tarnish plant bug, (suborder heteroptera, family miridae) was seen defoliating a non-lily plant next to a large cluster of L. PARRYI in Bear Canyon (Newman 1989a). In addition to damage of photosynthetic tissue, borers invade the inflorescence preventing seed development (Riggs 1989a, Warren et al. 1989). Boring insect damage in approximately 5% of the plants in Ramsey, Miller and Madera canyons was reported in 1980 and 1981 (Toolin 1982).