﻿Clermontiahanaulaensis (Campanulaceae, Lobelioideae), a new, critically endangered species from Maui, Hawaiian Islands

﻿Abstract Clermontiahanaulaensis H.Oppenheimer, Lorence & W.L.Wagner, sp. nov., a newly discovered, narrowly distributed endemic species, is herein described based on its morphological characteristics and illustrated with field photos and a line drawing. It is currently known only from the slopes of Hana‘ula, in Pōhākea Gulch, Mauna Kahālāwai, west Maui, Hawaiian Islands. It differs from all other species of Clermontia Gaudich. by the combination of its (2)3–4(–5) flowered inflorescence, violet colored perianth often suffused with creamy white streaks or sometimes creamy white with violet-purple irregular veins, (30)35–45(–50) mm long, perianth tube 15–25(–27) mm long, 9–10 mm wide, the lobes 20–26 mm long, (2–)3–3.5 mm wide, with petaloid calyx lobes 1/2–4/5 as long as the petals. A key to the Clermontia species and subspecies occurring on Maui is provided. Its habitat is described. Its conservation status is proposed as critically endangered (CR), and conservation efforts are discussed.


Introduction
The Hawaiian lobeliads (Campanulaceae, Lobelioideae) are the largest plant clade restricted to any archipelago, and originated from a single introduction ca. 13 mya (Givnish et al. 2009). In the most recent monograph of the genus Clermontia Gaudich., 22 species and nine non-autonymic subspecies were recognized (Lammers 1999). Clermontia occurs as terrestrial or epiphytic shrubs or small trees on the six largest Hawaiian Islands from 150 m to 2100 m elevation in mesophytic to wet forests, cloud forests, bogs, and shrublands. Lammers (1995) argued, based on a morphological analysis of relationships within Clermontia, that the genus arose on Maui rather than Kaua'i or older islands. However, Givnish et al. (2013) used plastid and nuclear DNA sequence data to show that Clermontia did arise on Kaua'i or an older island and largely obeyed the progression rule, with inferred inter-island dispersal events largely down the Hawaiian chain, from older to younger islands. Kaua'i, the oldest current high island is estimated to be 5.1 mya, and has just a single species of Clermontia. Hawai'i Island on the other hand is <1 mya and is home to 11 species (Lammers 1995(Lammers , 1999. Maui has 14 taxa (not including C. hanaulaensis). One is presumed extinct (C. multiflora Hillebr.), while C. peleana Rock subsp. singuliflora (Rock) Lammers is apparently extirpated on east Maui, but is still extant on Hawai'i Island. Clermontia lindseyana Rock is included in this tally based on two specimens at the Bernice P. Bishop Museum Herbarium [BISH], although populations on the leeward slope of Haleakalā are probably not this taxon, but may instead be C. kakeana Meyen (R. Pender, pers. comm.). Therefore, additional studies are needed to determine if C. lindseyana occurs on Maui.
During the course of rare plant field work in the mountains of west Maui, plants belonging to the genus Clermontia were found across several small ridges and gullies on the slopes below Hana'ula, in the Pōhākea Gulch drainage basin (Fig. 1). Initially only eight mature individuals and several immature plants were observed. These plants were not assignable to any taxa known from Mauna Kahālāwai, or elsewhere on Maui. Three possibilities were considered.
The first possibility considered was that these plants represented natural hybrids. Three congeners occur sympatrically in the study area: the at-risk, but locally common C. arborescens (H. Mann) Hillebr. subsp. arborescens, endemic to Mauna Kahālāwai; C. grandiflora Gaudich. subsp. grandiflora, also endemic to Mauna Kahālāwai, but not considered to be rare; C. grandiflora subsp. munroi (H. St. John) Lammers, which is less common locally, but one of the most widespread of all taxa in the genus, occurring on east and west Maui, Lana'i, and Moloka'i. However, the newly discovered plants do not resemble any of these three taxa as might be expected if they are hybrids, as they differ in the height of the plants, size and texture of the flowers, the length of the peduncles, the number of flowers per inflorescence, and the relative length of the calyx lobes compared to the corolla, among other features. The discovery of many more plants during subsequent surveys tends to negate the possibility that these plants could be hybrids as there are nearly 100 mature individuals and several dozen immature plants and seedlings occurring over an area of 0.1 km 2 . The second possibility considered these plants to potentially represent a new distributional record for a species previously known from another island. However, there have been no new records documented for the genus Clermontia, and only a single new distributional record exists for Hawaiian Campanulaceae, namely Lobelia hillebrandii Rock (Lorence et al. 1995;Wagner et al. 1999Wagner et al. , 2005. This possibility was discarded after unsuccessful attempts to key it out using the keys in Wagner et al. (1999) and Lammers (1991). Furthermore, images were sent to botanists at the Hawai'i Branch of the Division of Forestry and Wildlife (DOFAW) who did not recognize the new plants (Josh VanDeMark and Lyman Perry pers. comm.). Finally, flowers of C. hanaulaensis were compared to verified images of other Clermontia taxa on Richard Pender's Flickr website (https:// flickr.com/photos/123604592@N07/) which also did not result in a match.
The final possibility remaining was that these plants represent a new, undescribed taxon. Although new species of Cyanea Gaudich. continue to be discovered and described (Oppenheimer and Lorence 2012;Spork-Koehler et al. 2015;Oppenheimer 2020), only one new subspecies of Clermontia has been described in the past 30 years (e.g., C. grandiflora subsp. maxima Lammers).
No new taxon at full specific rank has been described and accepted since C. lindseyana Rock was published in 1962(Rock 1962. Study of the material collected, herbarium vouchers, literature, field observations, and photographs revealed these plants are morphologically distinct from all other known taxa and represent a new species, herein described.

Taxonomic treatment
Distribution. Clermontia hanaulaensis is known only from a single population on west Maui in several small ridges and gullies on the slopes below Hana'ula, in the Pōhākea Gulch.
Soil is of typical basaltic origin derived from the original shield-building Wailuku Volcanic Series (Stearns and MacDonald 1942). The average annual rainfall is approximately 2700 mm. (Giambelluca et al. 1986).
Clermontia arborescens subsp. arborescens occurs sympatrically throughout the range of C. hanaulaensis, whereas C. grandiflora subsp. munroi is scattered although locally common, and C. grandiflora subsp. grandiflora occurs at the upper perimeter of the population. Even further away in much wetter habitat with annual rainfall above ca. 2900 mm are populations of C. micrantha (Hillebrand) Rock, while C. kakeana occurs in lower, drier areas with annual rainfall below ca. 2600 mm. These taxa are readily distinguished morphologically from C. hanaulaensis by the characters given in the key to the Maui species below.
Phenology. Clermontia hanaulaensis has been observed to be flowering from July through September, with fruit maturing from August through October.

Affinities
Hillebrand (1888) recognized 11 Clermontia species with six varieties, divided into two sections: Section Genuinae with calyx lobes as long as the corolla or a little shorter, and Section Clermontioideae, with the calyx lobes shorter and persistent. Rock (1919), in his monograph of the Hawaiian species of Tribe Lobelioideae, maintained both of Hillebrand's Sections and recognized 23 species with four infraspecies. These two divisions are consistent with later treatments, but Section Genuinae was nomenclaturally corrected to Section Clermontia by Lammers (1991). In the most recent monograph of the genus, Lammers (1991) divided the genus into two sections, each with three series for a total of six taxonomic subdivisions.
A re-analysis of Lammer's morphological data by Givnish et al. (2013) shows that the two sections are not supported even by morphology when properly analyzed, with almost all resolution disappearing in the unweighted strict consensus tree. Furthermore, their analysis of plastid and nuclear sequence data demonstrates that petaloid sepals -almost certainly the product of a single homeotic mutation (Hofer et al. 2012;Givnish et al. 2013) -do not define a clade in Clermontia, with maximum-parsimony and Bayesian inference both showing multiple origins and losses of the trait. Similarly, Pender (2013) concluded that these entities were not fully supported by cpDNA. Givnish et al. (2013) demonstrated that the molecular data place Clermontia fauriei H. Lév. from Kaua'i, and then C. persicifolia from West O'ahu sister to all other Clermontia species (mostly from Maui Nui and Hawaii), pointing to an initial adherence to the progression rule. Finally, they show that C. pyrularia Hillebr. is actually a species of Cyanea. Hunter (2018) used large amounts of cpDNA and nuclear sequence data to show that Clermontia s.s. is sister to the purple-fruited clade of Cyanea, and that the orange-fruited clade of Cyanea is sister to them both.
The corolla of C. hanaulaensis is weakly zygomorphic to almost rotate at anthesis with the lobes connate only in the basal half. Although Lammers (1991: 8-9) discussed perianth shape and uses this character throughout his key, these states are often not clear-cut or apparent in herbarium specimens and may depend on the stage at which the flowers were pressed. Photos of mature fresh flowers shows they are at least somewhat zygomorphic in virtually all Clermontia species [https:// flickr.com/photos/123604592@N07/], even those described by Lammers as being rotate. Field photos of C. hanaulaensis flowers show a transition from somewhat zygomorphic to almost rotate as the perianth lobes mature and spread (Fig. 3).
Because gene regions of this new species have not yet been sequenced or included in molecular-phylogenetic studies of Clermontia and additional work needs to be done regarding relationships within the genus, our inferences about putative interspecific relationships are based on morphological characters. Clermontia hanaulaensis resembles C. samuelii C.N. Forbes with two subspecies on east Maui (subsp. hanaensis (H. St. John) Lammers and subsp. samuelii). However, the latter species differs by its smaller leaves with the blade 5-11 cm long and sparsely to densely pubescent beneath, relatively larger flowers with a more strongly curved or arcuate perianth tube 20-38 mm long, shorter less spreading perianth lobes 10-20 mm long, and larger fruits 28-35 mm long and 15-18 mm wide. We have adapted the following key from Lammers (1999) to include C. hanaulaensis, which keys out in couplet 10. Clermontia lindseyana has been considered to occur on Maui based on two poor and hard to identify collections and is included in the key. It is likely that they represent C. kakeana, and if so, C. lindseyana would be endemic to Hawai'i Island and does not occur on Maui; nevertheless we have left it in the key.
Several atypical individuals were encountered among the population of Clermontia hanaulaensis. One individual has pure white outer and inner perianths that are also shorter and wider than typical for C. hanaulaensis (Fig. 3 F, H. Oppenheimer & K. Severson H82101 BISH, PTBG, US). Two others have purple-pink and white perianth lobes that are shorter and wider than typical for C. hanaulaensis (Fig. 3D, E), and in one the calyx tube is slightly separated from the corolla tube (Fig. 3D). These variants were excluded from the circumscription of C. hanaulaensis on the basis of their morphology, although we mention them as they may represent hybridization between C. hanaulaensis and other sympatric species or between other species in this area. Further study is needed, but is beyond the scope of this paper. Inflorescence 7

Conservation status
Clermontia hanaulaensis should be considered Critically Endangered (CR) due to its limited range and low population numbers (ca. 120-130 individuals), assumed loss and/or decline of most or all of its avian pollinators and dispersal agents, threats such as landslides and treefall, herbivory by alien slugs ( to the subsequent alien plant invasion. The progressive upslope spread of axis deer (Axis axis) from areas below the population of C. hanaulaensis is of growing concern. Stochastic events such as a hurricane, as well as landslides and treefalls, have the potential to gravely impact this species. Fire is also a threat, although the cumulative effects from climate change are presently unknown.
When evaluated using the World Conservation Union (IUCN) criteria for endangerment (IUCN 2019), Clermontia hanaulaensis falls into the Critically Endangered (CR) category, which designates species facing a very high risk of extinction in the wild. The CR designation is indicated when any of the criteria A to E are met. Both Criterion B1 (Extent of Occurrence or EOO) and B2 (Area of Occupancy or AOO) are met with an EOO of less than 100 km 2 and an AOO of less than 10 km 2 , occurring in a single population of less than the minimum, single 2 km 2 grid. Criteria A & C address an anticipated decline in population for which there is no data. However, decline is inferred by the lack of robust demographics, with many fewer immature and seedling stages observed compared to mature, reproductive individuals. Criterion D reflects a very limited distribution, currently calculated as a single population of 0.1 km 2 . No quantitative analysis predicting the likelihood of extinction (Criterion E) was conducted. The alphanumeric formula CR B1a, b(v)+B2a,b(iii,v)+C1 represents the current status under IUCN guidelines.
Furthermore, Clermontia hanaulaensis should be considered by the US Fish and Wildlife Service as a Candidate for listing as Endangered under the Endangered Species Act of 1973, and a Recovery Plan written, funded, and implemented.

Conservation efforts
It is remarkable that this species occurs in a relatively accessible area (no helicopter support is necessary) that has been botanized reasonably well over the past 150 years. Mann & Brigham first collected Phlegmariurus mannii (Hillebr.) W.H. Wagner "on the mountains above Ma'alaea bay" (Hillebrand 1888). Degener collected the type of Clermontia furcata F.E. Wimmer (now a synonym of C. arborescens subsp. arborescens) "mauka of McGregor" in 1952 (Degener 1956). Field work was carried out in this area for at least three decades to varying degrees by Robert Hobdy, Steven Perlman, Kenneth Wood, as well as the first author. Botanical survey work has taken place on adjacent ridges, and further search effort is warranted and planned. This species may have the most restricted range of any taxon in the genus, occupying an area of only 0.1 km 2 .
The Maui Invasive Species Committee (MISC) has been working to control the incipient Cortaderia jubata (Lemoine ex Carrière) Stapf infestation in adjacent areas. The region formerly had feral pigs (Sus scrofa) that the previous landowner, Wailuku Water Co., along with the Mauna Kahālāwai Watershed Partnership (MKWP), formerly West Maui Mountains Watershed Partnership, has successfully controlled through strategic fencing, at least for the time being. Domestic and escaped cattle grazing on adjacent lands have occasionally entered the study area; this threat has also been mitigated, and cattle are no longer on adjacent lands. Axis deer (Axis axis) range is pushing uphill, and the Division of Forestry and Wildlife (DOFAW) of the Hawai'i Department of Land and Natural Resources (DLNR) and MKWP have begun to renovate an existing pig fence to exclude deer as well as implementing other strategies to limit the size of the herd (Lance DeSilva, DOFAW, pers. comm.).
While this new species exceeds the threshold of 50 wild individuals required to be a target of the Plant Extinction Prevention Program (PEPP), the population will continue to be monitored in case it undergoes a decline. More than 80 mature individual plants have been mapped with GPS, flagged and tagged to assist monitoring and ensure genetic representation ex situ. Almost 26,000 seeds from 20 individual plants have already been collected and are in storage at the Lyon Arboretum Seed Storage Laboratory in Honolulu. The PEP Program (http://www.pepphi.org/) strives to collect seeds or cuttings from every individual plant on the USFWS endangered species list, with ex situ seed storage, propagation of nursery stock, restoration outplantings into appropriate habitat, and living collections being the main conservation goals.

Data availability
All of the data that support the findings of this study are available in the main text or Supplementary Information.