Lucky morning glory, Calystegia felix (Convolvulaceae): a new species from Southern California, with notes on the historical ecology of the Chino ciénega belt

Abstract A new morning glory, Calystegia felixProvance & A.C. Sanders sp. nov. (Convolvulaceae), is described from the Los Angeles, San Gabriel, and Santa Ana River basins. Historical collections of the species, which prior to 2011 had not been seen alive in 94 years, have been misidentified as Calystegia sepium (L.) R. Br. subsp. binghamiae (Greene) Brummitt. The undescribed species was rediscovered in the City of Chino in April of 2011, a few miles north of the location where the most recent previous collection had been made by I. M. Johnston in 1917. The plants were found just prior to their likely destruction by grading and trenching for an underground power line. Intensive searches have resulted in the discovery of five additional occurrences, all of them in the City of Chino. Calystegia felix is at high risk of soon becoming extinct in the wild. All of the known extant occurrences are associated with well-watered landscaping on recently completed industrial, commercial, and residential developments. Every known living occurrence is within the limits of a ciénega belt, which is now mostly historical. Otherwise, the new species is only known only from collections made around the turn of the 20th century in what are now heavily urbanized areas, including one from South Los Angeles and another from Pico Rivera in Los Angeles County. Calystegia felix lacks the large bracts that immediately subtend, and enclose the calyx, which are always present in members of the Calystegia sepium complex. Affinities to Calystegia felix are found among other western US species with graduated sepals and small, often somewhat remote bracts. We discuss the enduring confusion between Calystegia felix and Calystegia sepium subsp. binghamiae, and differentiate the new species from some of its more likely relatives. The taxonomic treatment is supplemented by photos of herbarium specimens and living plants. We also discuss the ecological setting of Chino’s ciénega belt, which was a mosaic of palustrine wetlands.


Introduction
Calystegia R.Br. is a genus of about twenty-fi ve species, having a worldwide distribution in temperate zones. Its center of diversity is California, where twelve native species, and thirteen additional native subtaxa, are recognized in the latest Jepson Manual (Brummitt 2012). Although similar to Convolvulus L., several characters have been used to diff erentiate Calystegia from Convolvulus , including pantoporate pollen, oblong stigma pairs with blunt apices, and a unilocular ovary (Lewis and Oliver 1965). Despite the apparent morphological diff erences, molecular phylogenetic studies in 2002 (Stefanovic et al. 2002) and 2007 (Carine et al. 2007) suggested that Calystegia was nested within Convolvulus , making Convolvulus paraphyletic with respect to Calystegia . However, the studies supported a monophyletic Calystegia , but sampled very few members of the genus. Furthermore, the studies did not include species with graduated sepals and small, remote bracts.
In 2011, the fi rst author discovered a morning glory in Chino, California, that could not be diff erentiated from collections determined by the late R. K. Brummitt as Calystegia sepium (L.) R. Br. subsp. binghamiae (Greene) Brummitt, or "Calystegia binghamiae " (an unpublished name that he had written on one of the specimens years before). Until rediscovered in 2011, this taxon had been widely considered extinct (Brummitt 2012), as the most recently collected specimen authenticated by Brummitt was a specimen collected at Chino Creek in 1917( I.M. Johnston 1274. Th e site of the rediscovery was along a public walkway in a Southern California Edison (SCE) right-of-way in the City of Chino, roughly four to fi ve miles north of Johnston's collection locality. Th is morning glory was confi ned to an irrigated open-bottom planter at ground level on native soil.
Th e rediscovery was followed by elevation of C. sepium subsp. binghamiae to species rank under Calystegia ) because a reevaluation of the material indicated to them that the taxon warranted species rank. Commendably, their paper drew attention to the astonishing reappearance of a rare taxon in Chino that had been unseen for 94 years. It also completed the lectotypifi cation of Convolvulus binghamiae Greene, which had been initiated by Jepson (1939) when he made an earlier combination, Convolvulus sepium var. binghamiae (Greene) Jepson. Brummitt et al. (2012) noted that more eastern specimens of Calystegia binghamiae had more linear to narrowly elliptic bracteoles than were observed in western and more northern populations, and that sepals were at least sometimes inserted signifi cantly below the sepals. Th ey also recognized diff erences in leaf characters, particularly with regard to basal lobe size and shape. After having received both the lectotype and high-resolution digital images of all of the other known original material (except an isolectotype putatively held at F) for closer examination, we are now certain that the Chino material represents a distinct taxon that does not bear a close relationship to Convolvulus binghamiae Greene.
Th e six recent Calystegia locations from Chino are a new species. Th is new species is also known from three historical collections, including one each from South Los Angeles, Pico Rivera, and Chino. Calystegia felix Provance & A.C. Sanders, sp. nov., lacks the large bracts that immediately subtend, and clasp the calyx, as are always present in members of the C. sepium complex (Brown et al. 2009). We suspect that C. felix is more closely related to species of Calystegia endemic to the western US that possess graduated sepals, and have relatively small and sometimes remote bracts. Based on data gained from personal observations, herbarium collections, the early literature, and old maps, we think that C. felix is restricted to ciénega-wetland complexes in Southern California's alluvial basins. Th e extant occurrences in Chino must represent either plants that germinated from latent seed banks or are resprouts from the roots of plants that persisted in fi elds when this was an agricultural area. Either way, they apparently reappeared following the introduction of landscaping practices that have brought about "moist ground" conditions, similar to those that were historically present.

Methods
Recent Calystegia specimens from Chino were compared with Calystegia and Convolvulus collections at RSA-POM and UCR, as well as a selection of specimens from CAS and UC-JEPS, including the lectotype of Convolvulus binghamiae . We also obtained high-resolution digital images of Convolvulus binghamiae original material held at G-NDG. Scores of Convolvulus and Calystegia specimen images were evaluated for their relevance to the present study. We obtained images of relevant specimens held at the following herbaria: CAS, DS, E, K, LSU, NA, NY, P, SOC, UC-JEPS, US and WWB. Specimens with little immediate relevance are not listed in the appendices; however, regardless of ultimate relevance, all image sources and their source herbaria, are provided in Appendix I. Calystegia felix specimens that were examined are cited in the taxonomic discussion. All specimens of C. sepium examined are cited in Appendix II. Species of Calystegia that we think could be most easily confused with Calystegia felix are compared across a number of characters ( Table 1). Specimens that we examined of the species included in that table are cited in Appendix III. Clarifi cations added to specimen citations appear in brackets. Th e precisions of the reported geographical coordinates were reduced to ± 300 m for the C. felix collections. Only specimens and specimen images that were seen by the authors are cited in the appendices, except in the case of a putative sheet of Convolvulus binghamiae at F ( Mrs. R.F. Bingham s.n. ) cited in the original description, and one specimen of C. felix ( J.M. Wood et al. 4092 ). Th e herbarium code is followed by the word "image" when only an image was examined. Measurements were obtained from specimens conventionally, or from images using Meander V 2.3 (Dixon and Coventry 2008, available at http://www.fastforwardsw. com/products/ occidentalis by its narrowly oblong, 2.5-5 mm wide sepals (versus oblong to oblongovate, 6-9 mm wide sepals in C. occidentalis subsp. occidentalis ), narrower corolla tube (basally) 4-5.9 mm wide measured at the most proximal visible point (versus 6-9 mm in C. occidentalis subsp. occidentalis ), an ovary that is glabrous on inside walls (versus a silky hairy vestiture inside of the ovary in C. occidentalis subsp. occidentalis ), and larger, 45-122 mm long, 30-96 mm wide, oblong-ovate to broadly ovate leaves subtending the peduncles (versus smaller leaves subtending the peduncle, 32-51 mm long, 33-66 mm wide, narrowly triangular to broadly ovate), and short, rounded, barely divergent to parallel basal lobes, or leaves that are nearly truncate at the base (versus leaves with divergent basal lobes of varying length that are 2-lobed to bipartite). Description. Semi-herbaceous perennial vines, senescing in October, though with some stems and leaves persisting through winter. Aerial stems 1-3 m long, from shallow, creeping rhizomes and stolons (Fig. 2), climbing and twining, or clambering across shrubs, branching frequently, terete, with nonobvious longitudinal ridges, slender, tough and wiry, glabrous to sparsely hairy, in life dull grayish pink to light green with a rosy cast. Leaves alternate, membranaceous to chartaceus, glabrous to sparsely hairy, bicolored when mature green above, paler below, relatively fl at and not fold- ing along the midrib, but sometimes the basal half of the lamina slightly involute, and often having the basal lobes abruptly turned upward. Petioles on climbing stems 0.3-0.5 × length of lamina, e.g. about 14-61 mm long, but often longer relative to lamina length on emergent leaves; lamina of climbing stems 45-115(-122) mm long, 30-80(-96) mm wide, oblong-ovate to broadly ovate, but narrowly oblong on the sterile branchlets and on stems distal to the fl owering axils, base cordate, with short, rounded, parallel or barely diverging basal lobes, sometimes essentially without basal lobes and nearly truncate, apex obtusely rounded, sometimes subacute, minutely apiculate; emergent leaves from rhizomes and on trailing stems variable in shape, but usually broadly oblong to oval or orbicular, sagittate, with short lobes, or lobeless and rounded to the petiole, apex broadly rounded; lamina venation obscurely pinnate, but with 2-4 lateral veins from the base. Infl orescences axillary, fl owers usually solitary, rarely 2-3(-4)-fl owered; pedicels 1-30 mm long, peduncles 18-63 mm long; bracts 2, attached (1-)2-3(-4) mm below the calyx, ascending, subopposite, 5-14 mm long, 1-2.5(-3.5) mm wide, narrowly elliptic to narrowly oblanceolate, obtusely pointed, ± fl at, with a raised midvein, glabrous to scantly puberulent. Flowers perfect; sepals 5, entire, graduated, narrowly oblong to lanceovate, green with a rosy blush, short-ciliate, inner sepals 11-15 mm long, 3.5-4 mm wide, the lower portion tightly appressed to mature fruit, outer sepals 8-11 mm long, 2.5-5 mm wide, apices ± acutely rounded, mucronulate; corolla funnelform, 27-45 mm long, base of visible tube 4-5.9 mm wide, white (sometimes appearing light yellow in herbarium specimens), with 5 externally pigmented interplicae (midpetaline bands or longitudinal stripes), these very light-yellow, more rarely reddish-purple (Fig. 3), glabrous externally, or rarely, conspicuously hairy adjacent to pleats in the basal third of the corolla, the hairs yellowish, lobes 5, very short, each with a concentrated area of minute hairs along the margin; stamens 5, equal; fi laments 18-21 mm long, fused to the corolla tube ± 7-9 mm of that length, glandular hairy along the proximal margins; anthers 4-4.5 mm long, white, barely reaching the base of the stigmas; pistil glabrous both internally an externally; style 16-21 mm long, glabrous, or with a few glandular hairs near the base; stigmas 2, cylindrical, ± 3 mm long, asymmetric, with one axially oriented, and the other ascending; nectary crenate-coronoid. Pollen white to cream, with circular perforations discernible at 60 X. Fruit dry capsule, indehiscent to tardily dehiscent from tip to base, globose, 9-10 mm in diameter, glabrous or obscurely minutely pubescent apically. Seeds 1-4 per capsule, ca. 4 mm in height and 3.5-4 mm in width, ± angular-ovoid, and depending on the number of developing seeds, nearly black to dark brown and tan-speckled, hilar region purplish, fi nely granular.
Distribution. Calystegia felix is endemic to the inland basins of the Los Angeles, San Gabriel, and Santa Ana river watersheds in Southern California, at between 40 and 208 meters elevation. Th e species has not been seen in Los Angeles County since 1902. Six occurrences are known, all of them in the City of Chino, in San Bernardino County (Fig 4). Th e occurrences have a spatial separation ranging from 0.3-2 km. Th e easternmost occurrence is just west of Euclid Avenue, close to Chino's border with Ontario and Eastvale. Th e westernmost occurrences are on alluvial terraces above Chino Creek, coming within several meters of the City of Chino Hills.
Phenology. Flowering begins in late March, and is heavy until early August, with fl owering thereafter decreasing through late September. In 2011 and early 2012, infl orescences on the only plants known at that time had solitary fl owers. During mid-May of 2012, infl orescences at that site were observed to be two or three-fl owered, and rarely solitary. It is not certain whether fl ower number increased as the season progressed, or if fl owers during the later visit were originating on vines of a diff erent genet. Only solitary fl owers were seen at the sites discovered in 2013. Ripe seeds have been collected from early June until late October. Fruit with small holes indicative of seed predation by bruchid beetles have been found (Provance, pers. obs.). Small, senesced, nodding, sterile, apetalous fl owers, mostly near ground level, have recently been noticed on some plants. We observed similar fl owers on herbarium specimens of a few other species of Calystegia. It is unknown if these fl owers are apetalous developmentally, or if the corollas were lost to insect predation. More in-depth study of this condition is needed.  Discussion. Taxonomy: Although arguments to maintain Calystegia have been weakened by recent molecular studies, we describe this new species as such, pending molecular phylogenetic studies that sample more thoroughly across both Calystegia and Convolvulus . Although their fl owers are not similar, it is noteworthy that few vegetative characters seem to separate C. felix from the weed Convolvulus arvensis L. Th e only vegetative feature we currently know that can reliably be used to tell these taxa apart is the cross section of the stem, which is angular in Convolvulus arvensis , and terete with weak longitudinal ridges in C. felix . Th ere may be diff erences in leaf venation, but that will require additional study. Unfortunately, Convolvulus arvensis is abundant throughout the Chino area, and occurs at several of the C. felix sites.
While a defi nitive treatment of the entire C. sepium complex has not been published, the best defi ning features of this group are the large bracts which immediately subtend, and often enclose the calyx, and have conspicuously netted venation. It is a taxonomically diffi cult complex that may include over twenty C. sepium subtaxa, some additional closely related species, and their subtaxa (Stace 1961, Brown et al. 2009). So defi ned, all of the original material for Convolvulus binghamiae , including the lectotype at UC (Brummitt et al. 2012), is clearly referable to this complex. Th e epithet binghamiae is sometimes applied to specimens from Southern California with clasping bracts that only partly cover the sepals, and have leaves with barely divergent basal lobes, but that are otherwise inseparable from C. sepium . All of the original material of Convolvulus binghamiae is from a salt marsh that once occurred in Santa Barbara. Th e lectotype has only one fl ower (Fig. 5), which has a bract that may be the smallest found on any of the original material. Nonetheless, the corolla has a broad base, as seen in all members of the sepium complex. Moreover, leaves from the same sheet ( Fig. 5 and Fig. 6) are consistent with many specimens attributable to the Calystegia sepium complex. Th e bracts among the original material range from 7 to 13.8 mm long and 3.1 to 9.1 mm wide, and in their fully developed state are broadly lanceolate to broadly ovate. Infl orescence bracts in C. felix have a similar range in length, but are much narrower at 1 to 3.5 mm in width, and usually lack a conspicuous network of veins. Most bracts in the Convolvulus binghamiae original material are in every sense typical of the C. sepium complex. Interestingly, some of the largest bracts are associated with fl ower buds: in one case (Fig. 7), some of the bracts of the fl ower buds are larger than the bracts of the open fl ower on the same sheet. Finally, the corolla tubes (measured at the base of the sepal lobes) in the C. binghamiae original material are over 8 mm wide. In C. felix the lower tube of the corolla is narrow, ranging from 4 mm to about 6 mm in width. Calystegia felix is clearly not part of the Calystegia sepium complex, and represents a new, unrelated, and previously undescribed taxon.
Specimens of C. felix were included in Convolvulus binghamiae by Davidson in his list of new records for Los Angeles (1909) and by Davidson & Moxley in their fl ora of Southern California (1923). When Jepson made the combination Convolvulus sepium var. binghamiae (Greene) Jepson, he was very particular in his application of the name, stating, "Santa Barbara; a distinct localized variety, rarely collected". Under Convolvulus sepium var. pubescens , Jepson (1939) cites a Chino specimen ( Condit s.n. ) that we have seen, and which is equivalent to material typically ascribed to C. sepium subsp. limnophila in Southern California (and elsewhere). While it is possible that Jepson had In 1945, Abrams annotated one sheet of Johnston's C. felix collection at RSA as Convolvulus binghamiae , probably while preparing his Illustrated Flora (1951). Th e illustration of Convolvulus binghamiae in this fl ora seems to be C. felix , which is incongruent with his treatment, since the geographic distribution given by Abrams for Convolvulus binghamiae excludes all extant and historic occurrences of C. felix. Brummitt (1993) applied Calystegia sepium subsp. binghamiae (Greene) Brummitt, to plants of the northern and central South Coast between sea level and 20 meters elevation, which excludes collections from Chino. In Brummitt et al. (2012), the author's recognize the similarity of material we refer to C. felix to the illustration in Abrams fl ora, noting, "A good illustration of the latter may be seen in Abrams (e.g., Fig. 3855, 1951)". However, they propose that the drawing represents a phenotypic variant of their proposed Calystegia binghamiae (Brummitt) Brummitt. Th e misidentifi ed collections at RSA and the illustration in Abrams of what was actually an undescribed species seems to have infl uenced the search image of at least some local botanists attempting to rediscover C. sepium subsp. binghamiae .
Philip A. Munz annotated a Johnston collection at RSA as Convolvulus purpuratus Greene in 1931, thus clearly including C. felix in his concept of Convolvulus purpura- tus . In his Southern California Manual (Munz 1935) he listed Chino as a locality for this species. Among many diff erences, Calystegia purpurata (Greene) Brummitt subsp. purpurata can be readily separated from C. felix by its equal, evenly inserted sepals.
Th e number of fl owers per infl orescence, corolla pigmentation, external corolla vestiture, and the vestiture of leaves and stems vary in C. felix. Whether this variation is infl uenced more by genetics or environmental factors remains to be studied. Heterophylly is profound in C. felix , and generally manifests as narrower lamina on sterile twining stems, instead of the larger ovate to oblong-ovate leaves of reproductively active stems. Th ere seems to be a tendency towards rounder leaves with longer petioles on emergent stems and sometimes trailing stems. Calystegia felix is similar to other species in the genus with small, somewhat remote bracts, and graduated sepals. Several morphological characters are used to compare four of those species with C. felix (Table 1). Leaf parameters alone are often insuffi cient for the identifi cation of Calystegia , but fortunately, several other characters in addition to leaf shape, diff erentiate C. felix from other species of Calystegia .
At fi rst glance, C. felix looks most similar to C. occidentalis (Gray) Brummitt subsp. occidentalis , since both taxa have a similar clambering or climbing habit, similar bracts inserted approximately the same distance below the calyx, and potentially produce multiple fl owers in infl orescences. However, C. felix diff ers from C. occidentalis by its narrower sepals, narrower corolla tube, internally glabrous ovary, and larger oblongovate to broadly ovate leaves. Th e leaves subtending peduncles of C. felix have short, rounded, barely diverging to parallel basal lobes. Sometimes, C. felix leaves are nearly truncate at the base. Th is easily diff erentiates C. felix from C. occidentalis , which has lamina basal lobes that are of varying length, but divergent, and usually 2-lobed or bipartite. Calystegia felix also looks like C. subacaulis Hook. & Arn subsp. episcopalis Brummitt. Both taxa have slender, but tough and wiry stems, corolla tubes that narrow toward the base, narrow sepals, and an ovary that is glabrous both internally and externally. It diff ers from C. subacaulis subsp. episcopalis by its strong climbing habit, and much larger leaves that diff er considerably in basal lobe morphology.
Th e similarities between C. felix and C. subacaulis subsp. episcopalis tend to be less readily apparent than the similarities between C. felix and C. occidentalis subsp. occidentalis. However, the characters shared seem not to be widespread in the genus. For example, while ovaries of C. felix sometimes have a small number of minute hairs toward the apex, they are essentially glabrous externally. Th ey are also glabrous internally. Th ough we have had only one specimen of C. subacaulis subsp. episcopalis upon which we have been able to conduct detailed fl ower dissections ( F. Bowcutt 2163 [UCR]), we are especially intrigued by the ovaries of this collection, which are glabrous both internally and externally. We have seen this combination of characters only in C. felix , and similarities such as these might indicate that the two taxa are more closely related than their superfi cial appearances suggest.
Ecology: Th e six known occurrences are associated with somewhat poorly drained alkali silt loam (SoilWeb 2013), on a fl oodplain with an average slope of just over 1% (Lewis Publishing Company 1890, SoilWeb 2013. Th e local soils have developed primarily from the accumulated granitic alluvium that was washed out of the San Gabriel Mountains during episodic fl ood events (Hilgard 1902). Historically, there were a number of springs near the C. felix occurrences, and the springs of the plains and basins and their accompanying vegetation, typically marshland and wet meadows, were known as ciénegas (Schuyler 1880, Mendenhall 1908. In Southern California the use of the word ciénega always implied the presence of a spring, unlike in some other parts of the American Southwest (e.g. Hendrickson and Minckley 1985). In the earliest known general description of ciénegas in the Chino Basin, Schuyler (1880) emphasized that ciénega was the only word commonly used to indicate its springs and associated habitat. In the Chino Basin, there were two (Hall 1888a(Hall , 1888b or three (Tait 1911) main groups of ciénegas located a very short distance west to southwest, south, and southeast of Chino (Tait 1911). Th e perimeter of the Chino Artesian Spring Belt was roughly triangular, and its location in the current landscape is easily derived from the early maps. Th e east and west vertices were near the Chino Creek and Mill Creek emergences respectively. A third vertex would be near the south side of Prado Basin. Th ese boundaries coincide well with the historical limits of "moist land" as mapped earlier by H. B. Martin (1887Martin ( -1889. Mendenhall (1908) estimated the area of the artesian belt and associated moist soil as 23 sq. miles prior to 1904. Various aspects of the hydrology and geology of the Chino Basin ciénegas have been summarized (e.g. Truman 1874, Schuyler 1880, Hall 1888b, Lewis Publishing Company 1890, Shinn 1898, Mendenhall 1905, Hilgard and Loughridge 1906, Troxell 1957. Historically, the water table in the vicinity of the artesian spring belt was 6-35 feet below ground (Lewis Publishing Company 1890) Th e soils within the spring belt, which are largely alkali silt loams, retained moisture throughout much of the year, and as a consequence were extraordinarily important to Southern California agriculture (e.g. Peff er 1894, Shinn 1898, Nelson 1917. Based on soil maps, four of the C. felix occurrences are on Chino silt loam. Both of the occurrences that are not on Chino Silt Loam, one on Grangeville fi ne sandy loam, the other on Hilmar loamy fi ne sand, are less than 30 feet from Chino silt loam according to soil maps (SoilWeb 2013). While Calystegia felix occurrences seem to be strongly associated with Chino silt loam; an analysis of soil at occupied sites has not been performed.
Although seeds and rhizomes can be moved around in many ways, we contend that invoking accidental transport of stem fragments or seed by humans is not the most parsimonious explanation for the presence of Calystegia felix in the City of Chino, since the species is known nowhere else. While we have no direct proof, we think the recently discovered Calystegia felix populations represent plants that have emerged from latent, long-lived seed banks or roots following a return to "moist soil" conditions (Figs 8,9), similar to those in the historical record. Buried seeds of C. sepium have retained high levels of viability after 39 years (Bond et al. 2007), and C. felix may have similar longevity. If changes in soil moisture regimes are occurring (i.e.becoming wetter), horticultural practices within the urban environment are likely the cause. We have not investigated soil moisture in Chino experimentally, but we observed an apparent moisture gradient. Th e success we have had locating new occurrences of this rare plant in developed areas contrasts sharply with our failure to locate occurrences in undeveloped visually drier areas. While we are not sure of the signifi cance at this point, it seems noteworthy that each of the sites currently supporting Calystegia felix were, based on aerial images (Google Earth V.2.1.6014b), completely stripped of their vegetation at some point between 4 and 11 years ago).
Historical information and early herbarium collections suggest that the Chino Basin originally had vegetation of wet meadow and alkali meadows dominated by Anemopsis californica (Nutt.) Hook. & Arn., with perennial grasses, such as Elymus triticoides Buckley, Sporobolus airoides (Torr.) Torr. , and Distichlis spicata (L.) Greene, and herbs such as Trifolium willdenovii Spreng., Trifolium wormskioldii Lehm., and Helianthus annuus L. In addition, there were small bodies of open water, alkali and freshwater marshes, alkali scrub, alkali grassland, alkali playa, moist stream banks, and willow thickets. Th ere were also phreatophytic woodland communities of Salix , Populus , and Platanus racemosa Nutt. (Truman 1874). Th e spring-belt wetlands were collectively referred to as "ciénega-lands" (e.g. Hilgard 1889). Common sunfl ower ( Helianthus annuus ) is frequently mentioned in the early literature as a common species on alkali soils in the Chino Basin, and was considered indigenous (e.g. Davy 1898). Th e topography just north of the Santa Ana River was reportedly hummocky (Nelson 1917), and may have been supportive of vernal pools. Most fi ne-scale relief in the basin has probably been lost to disking and grazing cattle, but the north part of the ciénega belt was reportedly rather smooth. Mendenhall (1905) commented: "Th e lands just above this [above the ciénega-lands] are fl at and often ill drained. Th e waters rising and evaporating here, under the infl uence of the eff ective southern sun, leave behind them their salt content, and thus alkali lands may result" Th us, much of the landscape represented a mosaic of ciénega and ciénega-creek associated palustrine communities. Th e historical natural vegetation of the City of Chino cannot easily be envisioned because of past and current development. For example, a satellite of the University of California Agricultural Experiment Station called the "Ten Acre Tract" used to be in Chino and experiments related to growing crops on alkali soil were conducted there. Th is property was described as being dominated by Anemopsis californica (Hilgard & Loughridge 1896), which indicates that it was likely alkali marsh . Th e Ten Acre Tract is now occupied by industrial buildings and offi ces. However, taxa highly indicative of alkaline marsh and alkaline meadow have persisted in unusual places. For instance, we documented a number of Anemopsis californica persisting in plantings of Hedera helix along a sidewalk in northeast Chino near the Ontario border, just within the mapped historical limits of moist ground. Th e following year we found Calystegia felix growing in a sidewalk planter across the street from the Anemopsis site, in similar urban landscaping. We think other remnants of the ciénega fl ora may persist in Chino.
Conservation : Calystegia felix is endemic to Southern California, extirpated in Los Angeles County, and now likely confi ned to the Chino Basin in San Bernardino County. It is doubtless at high risk of soon becoming extinct in the wild. Th is is due to hydrological changes in the Chino Basin, including the paving of streams, lowering of the water table, and loss of ciénega habitat, including vegetation associated with marshes, meadows, grasslands and alkaline playas; and encroaching commercial, industrial, residential developments, and public works projects. Large areas of habitat have already been transformed. Six extant occurrences are now known, with an estimated 200 ramets emerging in 2013 at a single location near Chino Creek. However, those plants likely represent clones, as do the about 50 ramets at the other sites. Based on there being few known populations, a limited overall distribution, and a small number of individuals in existence, we suggest a conservation status of Critically Endangered (CR). Upon discovering the plants along Edison Road, it was obvious they were in imminent danger of being destroyed by impending grading and trenching for the burial of high-voltage power lines. We initially thought that these plants represented a single clone, but two ramet-specifi c fl ower color morphs, seed production, and spatial separation of clusters of emergent stems, suggest that two or more genotypes are present. Over the short term, Rancho Santa Ana Botanical Gardens has been contracted to conduct ex-situ propagation of rhizomes from the Edison Road population, and a few other plants are being cultivated in private and institutional gardens.
Image databases, internet address , and source herbaria, of specimen images initially examined for this study.