A massive Critically Endangered cloud forest tree, Microcos rumpi (Grewiaceae) new to science from the Rumpi Hills, SW Region, Cameroon

We describe Microcos rumpi (Grewiaceae) a massive new species to science from the Rumpi Hills of SW Region Cameroon, a proposed Tropical Important Plant Area. Confined on current evidence to submontane forest, the species is threatened by expanding habitat clearance for farms and is assessed as Critically Endangered. A large tree, attaining 35 – 40 m height, and 80 cm trunk diam. at 1.3 m above ground, its biomass is calculated as in the range of 7 – 8 metric tonnes. It is the third tree species of the genus recorded from Cameroon and only the fourth recorded west of D.R. Congo. A key to these four species is presented. The concept of Microcos in Africa in relation to Grewia is discussed, and three new combinations are made, transferring three species names from Grewia to Microcos: Microcos louisii (R.Wilczek) Cheek, M. evrardii (R.Wilczek) Cheek and M. schmitzii (R.Wilczek) Cheek.


Introduction
In connection with the Cameroon Tropical Important Plant Areas programme (Darbyshire et al. 2017;Cheek, continuously updated), in May 2022, while clearing a backlog of specimens which had resulted from a survey of the Rumpi Hills and Nta Ali in SW Region Cameroon (Thomas 1995, supported by the last author), two specimens, Thomas 10426 and 10442 (both K!) that had been field-identified as Grewia coriacea Mast. (more correctly, see below, Microcos coriacea (Mast.) Burret), were found to be discordant with that species. Further checks against reference specimens in the Kew herbarium showed that not only did they differ from M. coriacea but from all other Microcos species in tropical Africa (see Results below). Accordingly, these Rumpi Hills specimens are here formally described and named as Microcos rumpi Cheek. Microcos Burm. ex L. (Linnaeus 1753) is a palaeotropical genus of about 78 species (Govaerts et al. continuously updated) based on M. paniculata Burm. ex L. (Linnaeus 1753) from Sri Lanka. Linnaeus (1767) later synonymised Microcos under Grewia L. However, the genus was resurrected by Burret (1926). Burret's authoritative revision (1926) of Tiliaceae sens. lat. presaged the former family's break-up into todays's Brownloideae/Brownlowiaceae, Tiliaceae sensu stricto/Tilioideae and Grewioideae/Sparrmanniaceae (Grewiaceae) including Microcos (Bayer et al. 1999;Bayer & Kubitzki 2003;Cheek in Heywood et al. 2007). Burret's was the last global treatment of Microcos (Burret 1926). He recognised 53 species, of which 19 were recorded from Africa and 34 in Asia. Of the 99 names in Microcos listed in IPNI (continuously updated), Govaerts et al. continuously updated) accept 78 species. The majority of these 78 are in SE Asia, but with 11 in Africa. Microcos is absent from the Neotropics and Madagascar. Within Africa, the species of Microcos can conveniently be divided into two groups based on habit: trees, often massive, such as that described in this paper, and climbers (rarely shrubs).
Illogically, while Microcos has been maintained as a separate genus from Grewia in Asia (e.g. Chung 2003Chung , 2006Chung et al. 2005a;Chung & Soepadmo 2011), the two genera have often been united under Grewia in Africa. For example, in one of the most recent Flora accounts of Grewia (including Microcos) for Africa, Whitehouse in Whitehouse et al. (2001) states " … Kirkup followed Burret in recognising Microcos as a distinct genus; this concept has also been followed in SE Asia. Although there are clear differences between Microcos and the other sections of Grewia, for consistency I am following the practice set by the other African floras, of not recognising…." This practice of retaining Microcos in Grewia is maintained widely today in Africa, for example by the excellent and essential African Plant Database (continuously updated). This also accounts for the discrepancy reported above in the numbers of Microcos species recognised from Africa between Burret (1926, 19 species) and Govaerts et al. (continuously updated,11 species). In fact, the two genera are readily separated as expressed in the key below, modified from that in Whitehouse (2001): Trees and climbers, rarely shrubs, of evergreen forest; stigmas entire; fruit unlobed; inflorescences terminal, sometimes axillary also, many-flowered………………………………………………………………………………………Microcos Shrubs, rarely trees, of bushland or woodland; stigmas lobed; fruit 4-lobed, rarely entire; inflorescences usually axillary or leaf-opposed, rarely terminal, usually few-flowered……….. …………………………………………………..Grewia According to the molecular analysis of Brunken & Muellner (2012), Microcos is not embedded in Grewia, neither are these two genera sisters, in fact they fall into distinctly separate clades. Additional characters for separating the two genera are found in the pollen, wood anatomy and in the leaf anatomy, particularly the epidermal cells (Chattaway 1934;Chung 2002;Chung et al. 2003Chung et al. , 2005b. Microcos was maintained in Bayer & Kubitzki (2003). However, in Africa, several species described in Grewia remain to be transferred formally to Microcos which is partly addressed here.
In contrast to Asia (see references above), the genus Microcos has been little studied in Africa, as evidenced by the fact that the first new name in African Microcos since 1926 was published in 2004 (M. barombiensis (K.Schum.) Cheek in Cheek et al. 2004: 414). Czarnecka et al. (2006) in their phytogeographic analysis of Microcos in Africa, also draw the conclusion that more taxonomic work is needed. In the course of matching the material described as new in this paper, it became clear that a revision of the genus for Africa is desirable to address specimen misidentifications and additional apparently undescribed species. It is hoped to address these problems in future.

Materials and Methods
Herbarium citations follow Index Herbariorum (Thiers continuously updated) and binomial authorities IPNI (continuously updated). Material was collected using the patrol method (e.g. Cheek & Cable 1997). Herbarium specimens were made in the field by the last author. Material of the suspected new species was compared morphologically with material of all other African Microcos (or Grewia sect. Microcos (L.) Wight & Arn.) principally at K and YA, but also using images of specimens available online, principally from BR, P and WAG. Burret's types of Microcos at B were destroyed by allied bombing in 1943 so it was not possible to consult them apart from images made available by F (anonymous reviewer pers. comm. Nov. 2022). The loss of types at B has necessitated that subsequent authors have often needed to select neotypes of Burret's names, e.g. Whitehouse (2001).
The conservation assessment was made using the categories and criteria of IUCN (2012). Herbarium material was examined with a Leica Wild M8 dissecting binocular microscope. This was fitted with an eyepiece graticule measuring in units of 0.025 mm at maximum magnification. The description terminology follows Beentje & Cheek (2003) and . The drawing was made with the same equipment using Leica 308700 camera lucida attachment.

Taxonomic Results
Here we formally transfer three more names from Grewia to Microcos for the reasons given in the discussion, and so that one of them can be referred to in the context of the taxonomic placement of the main subject of this paper, M. rumpi (see below). All three names are of taxa of D.R. Congo, published in what was intended to be a precursory account to the Flore de Congo Belge Tiliaceae treatment, although the last preceded the first by several months.
Previous year's stems with surface longitudinally furrowed and with raised, longitudinally elliptic, pale brown, lenticels 0.5 × 0.2 -0.35 mm. Leaves alternate, blades simple, entire, coriaceous, sub-bullate, drying mid brown, (ovate-) ovate-elliptic to elliptic-oblong (elliptic-obovate), (14.2 -) 15.1 -19 (-28.8) × (8 -) 9.2 -11 ( -11.8) cm, apex subacuminate or acumen short and broad, obtuse, 0.4 -0.8 (-0.9) × 0.4 -1.0 (-1.2) cm, base slightly asymmetric (broadly rounded-) truncate (-cordate), sinus c. 5 × 18 mm; abaxial surface pinnately nerved (not tri-or palmately nerved at base), lateral nerves prominent (7 -) 8 -9 (-10) on each side of the midrib, arising at c. 50° from the midrib, straight, and then arching gradually upwards, becoming parallel with the margin then uniting with the nerve above via tertiary nerves, forming looping inframarginal nerves 3 -4 mm from the margin in the distal half of the leaf. Domatia conspicuous, at junction of lateral nerves and midrib with a bright white dense tuft of hairs on each, the two tuft bases separated by c. 0.5 mm, the hairs 0.5 -1 mm long, the apices of the hairs of the two tufts interlocking (Fig. 1C). Tertiary nerves very prominent, scalariform, 14 (-17) between the basal and near basal secondary nerve pairs. Quaternary nerves barely detectable with the naked eye, reticulate, cells isodiametric, 0.75 -1.25 mm diam.; indumentum of dull white mainly 6 -8-armed stellate hairs 0.1 -0.2 mm diam., covering 10 -20 (-30%) of the surface; adaxial surface with nerves impressed, quaternary nerves invisible, glabrous; margin slightly revolute and thickened, entire. Sapling leaves (D. W. Thomas 10442, K) as mature, canopy leaves but papyraceous, lanceolate, 20 -21 × 8.9 -9.7 cm, acumen narrowly triangular 2 -2.3 × 0.6 -0.8 cm, base rounded-truncate, lateral nerves 9 -11 on each side of the midrib, quaternary nerves reticulate, conspicuous with the naked eye, finer nerves conspicuous with lens, also reticulate, cells c. 3 mm diam.; margin slightly sinuate, with veinlets terminating in minute purple glandular teeth, teeth 3 -4 mm apart, broadly obtuse, c. 0.1 × 0.15 mm, indumentum extremely sparse. Stipules lateral to petiole base, caducous, cicatrices white, subisodiametric, 0.7 -1 × 1 mm. Petiole cylindrical (0.8 -) 1 -1.4 (-1.5) × 0.2 -0.3 cm, indumentum as stem. Buds supra-axillary, inserted 0.5 -1 mm above the axil, narrowly ellipsoid 1.5 × 0.75 -1.2 mm, completely covered in grey stellate hairs. Inflorescence not seen, probably terminal, paniculate. Flowers (rehydrated from Thomas 10442) pre-anthetic buds obovoid-cylindric 7.5 -9 × 4 mm, pedicel 1.5 -2 × 1 mm, 5-ridged, ridges rounded, minutely puberulent. Sepals 5, valvate, divided to the base, oblong-spatulate, 6 -7 × 1.7 -2 mm, the proximal half narrowed, c.  1996). The site of the type specimen collection of the last, D. W. Thomas 10456 is given as "l.5 km W of Madie River ford". This species, in the identical habitat, with similar restricted range has been assessed for its global extinction risk status as threatened but is not yet published on iucnr edlist. org (July 2022). Microcos rumpi is here assessed as Critically Endangered since only a single site is known, area of occupation is calculated as 4 km 2 using the preferred IUCN grid cell size and extent of occurrence as the same. Plotting the grid reference of the two specimens on Google Earth shows that the site is outside the existing Rumpi Hills protected area, and examining historic imagery shows that between July 2008 and Jan. 2011, there is an increase in cleared areas of forest along the road c. 500 m to the west of the grid reference, and also clearing distant from the road. This justifies an assessment of CR B1ab(iii) + B2ab(iii). This distinctive, huge tree has not been found in surveys elsewhere in the Cameroon Highlands and adjacent areas (Cheek 1992;Cheek et al. 1996;Cable & Cheek 1998;Cheek et al. 2000;Maisels et al. 2000;Chapman & Chapman 2001;Cheek et al. 2004;Harvey et al. 2004;Cheek et al. 2006;Cheek et al. 2010;Harvey et al. 2010;Cheek et al. 2011). Therefore, it may indeed be endemic to the Rumpi Hills. ETYMOLOGY. Named as a noun in apposition for the Rumpi Hills in SW Region, Cameroon, to which this species is restricted on current evidence. VERNACULAR NAMES & USES. None are recorded. NOTES. The only known collections of Microcos rumpi appear to be connected with a transect made by Thomas (1995: 37 -39, table 5) in the Rumpi Hills with the support of the last author and the late Ferdinand Namata. In that transect, 607 stems above 10 cm dbh ("diameter at breast height", meaning trunk diameter at 1.3 m above ground level) were recorded in five size classes and allocated to 80 different species (Thomas 1995: 37 -39, table 5). The transect area was 1 Ha (1 km × 10 m) at elevation 1200 -1350 m alt., starting at the Madie River crossing between Dikome and Madie, just to the East of the RHFR. Hilltops were reported to be dominated by Santiria trimera (Burseraceae) a wellknown indicator species of submontane forest in Cameroon. Microcos rumpi (as "Grewia coriacea") occurred 10 times in total, 5 stems (c. 1% of stems) in the 10 -29 cm size classes, 1 stem in the 30 -49 cm class (c. 1.5%), and 4 stems (c. 15%) in the 50 -69 cm size class (Thomas 1995). Therefore, within the transect area, the species is not uncommon, since about 1 in every 60 stems above 10 cm dbh is M. rumpi. While the specimen location metadata relates to that given for the transect, the size class data given for the specimen (up to 80 cm dbh) exceeds that for any recorded in the transect, suggesting the specimen(s) were taken near to but outside the transect. Since only two stems among the 607 of the Rumpi Hills transect are placed in the 90+ cm size class (Thomas 1995), at 80 cm dbh (Thomas 10426), Microcos rumpi must be among the most massive trees in its Rumpi Hills cloud forest habitat consistent with the height given of 35 -40 m (Thomas 10426). Taking the average wood density of Microcos of 0.5 kg/dm 3 from Dryad (Zanne et al. 2009) in the absence of African data derived from Asian species of Microcos, together with the metadata cited above from Thomas 10426, the equation for estimating the biomass of trees from moist tropical forests (Chave et al. 2005) gave an above ground biomass of such a tree as approximately 6500 kg, plus 1400 kg below ground (equation from Sierra et al. 2001), that is 7.9 metric tonnes per tree. That such an immense organism could remain unknown to science until now is remarkable if not without precedent (e.g. Bayer & Dorr 1999).
At the time of specimen collection, it was evident from the specimen metadata that the collector, Duncan Thomas, queried whether these specimens represented Grewia coriacea (Microcos coriacea) since the trees were so much larger than those of the species encountered at lower altitudes. But by the time the report was written (Thomas 1995) the specimens are unambiguously attributed to Grewia coriacea (Microcos coriacea).
While Microcos rumpi is most likely to share a most recent common ancestor with one of the two other Cameroonian tree Microcos species, M. coriacea and M. magnifica (see above), it shares a remarkable character with one of the most common and widespread Congolese species, M. louisii. Both species have leaf domatia comprised of two separate clusters of white hairs which interlock at the apex (Fig. 1C). The two species share other similarities in leaf shape but differ in the features indicated in Table 1. In addition, M. louisii is mainly confined to seasonally flooded and swamp forests of the Congo basin and is not recorded from cloud (submontane) forest in Cameroon.

The Rumpi Hills
The Rumpi Hills Forest Reserve (RHFR) is situated 80 km N of Mt Cameroon, 50 km W of Bakossi Mts and 15 km SE of Korup National Park. The highest point is given as Mt Rata, at 1800 m (https:// en. wikip edia. org/ wiki/ Rumpi_ Hills). Data on the Rumpi Hills are scarce. The most authoritative appear to be a study on land-use change by Beckline et al. (2018). They state that the Rumpi Hills exhibits two seasons; the dry season from November to April and the rainy season from May to October with annual rainfall that ranges between 4027 -6368 mm. Mean monthly maximum temperature in the dry season is estimated at 31.8°C and 18.2°C during the rainy season. The relative humidity is high during most of the year with minimum monthly values ranging between 78% and 90%. Ethnic groups are the Ngolo, Bima, and Balue. The Rumpi Hills Forest Reserve is given as 455 km 2 and is kidney-shaped in outline. Google Earth imagery shows the forest canopy of RHFR to be largely intact, apart from oil palm plantations inside the far western boundary around the settlement of Ekumbako (4°55'41.47"N, 80°56'18.43"E). However, the boundary shown on Google Earth may be inaccurate and this settlement may fall outside the boundary. Despite the name, most of RHFR apart from inside the northern and eastern boundary, is not aligned with the upland area of the Rumpi Hills, which lie to the east. RHFR is below 600 m and mostly below 300 -400 m elevation. The highest settlement of the Rumpi Hills is Dikome Balue, the capital of the Rumpi Hills, which lies outside the RHFR. It is only 7 km S of the recorded location of Microcos rumpi. Smaller settlements, mostly above 1000 m alt. are scattered regularly through the hills proper, which occupy the area east and southeast of the RHFR. A recent study of changes in the Rumpi Hills between 2000 -2014 based on satellite images and, importantly, on-theground surveys in 14 settlements around RHFR and 200 land-use change plots, reported that during the 14-year period, dense forest dropped to 90.2% cover while settlements increased from 744.6 to 2148.8 hectares in 2014. Also, farmlands increased by 18.25% representing a change from 9,400 to 11,117 Ha (Beckline et al. 2018). Despite this, observation of Google Earth imagery dating from 2015 (accessed 17 July 2022) shows large areas of forest with intact canopy. The presence of the range-restricted endemics Ocotea ikonyokpe (see above), and, also outside the boundary of the RHFR, to the north, the point endemic Ledermanniella prasina J.J.Schenk & D.W. Thomas (Podostemaceae, Schenck & Thomas 2004), now joined by a third Rumpi Hills endemic species, Microcos rumpi (this paper) are indicators that the Rumpi Hills are a centre of plant diversity. That the numbers of recorded endemic plant species is so low compared with the neighbouring Bakossi Mts and Mt Kupe (Cheek et al. 2004) is undoubtedly down to the comparatively very low levels of botanical survey and identification in the Rumpi Hills. Undoubtedly many additional new species to science, including Rumpi Hill endemic species, will be brought to light if botanical collections and identifications are resumed in the Rumpi Hills while natural habitat remains.

Conclusion
It is important to uncover the existence of previously unknown plant species as soon as possible and to formally name them. Until this is done, they are invisible to science and the possibility of their being assessed for their conservation status and appearing on the IUCN Red List is greatly reduced (Cheek et al. 2020b), limiting the likelihood that they will be proposed for conservation measures, and that such measures will be accepted. Although there are exceptions (Cheek & Etuge 2009;Cheek et al. 2019a), most new plant species to science are highly range-restricted, making them almost automatically threatened (Cheek et al. 2020b).
Only 7.2% of the 369,000 flowering plant species (the number is disputed) known to science have been assessed on the IUCN Red List (Bachman et al. 2019;Nic Lughadha et al. 2016, 2017. However, the vast majority of plant species still lack assessments on the Red List (Nic Lughadha et al. 2020). Fortunately, Cameroon has a plant Red Data book (Onana & Cheek 2011), which details 815 threatened species, but it needs updating. Thanks to the Global Tree Assessment (BGCI 2021) many of the world's tree species have now been assessed. The State of the World's Trees concluded that the highest proportion of threatened tree species is found in Tropical Africa, and that Cameroon has the highest number (414) of threatened tree species of all tropical African countries (BGCI 2021). This will be further increased by the addition of Microcos rumpi.
Concerns about global plant species extinctions are increasing as the biodiversity crisis continues. In Cameroon, the lowland forest species Oxygyne triandra Schltr. and Afrothismia pachyantha Schltr. have been considered extinct for some years (Cheek & Williams 1999;Cheek et al. 2018b;Cheek et al. 2019b). Similarly, Pseudohydrosme bogneri Cheek & Moxon-Holt and P. buettneri Engl. are now considered extinct in lowland forest in neighbouring Gabon (Moxon-Holt & Cheek 2020; Cheek et al. 2021b). However, submontane (cloud) forest species are now also being recorded as extinct in Cameroon, such as the well-documented case of Vepris bali Cheek et al. 2018c), and the only recently discovered Monanthotaxis bali Cheek (Annonaceae, Cheek et al. 2023). Global species extinctions are being recorded from across Africa, from West (e.g.
If such extinctions are to cease, or more realistically, to be slowed, improved conservation prioritisation programmes are needed to firstly determine the most important plant areas for conservation (Darbyshire et al. 2017) and secondly to implement protection with local communities and authorities through well-drawn up management plans and resourcing. Cultivation and seed-banking of species at risk of extinction, if feasible, are important fall-back strategies, but conservation of species in their natural habitat must be the first priority.
Happily, the Rumpi Hills were recently designated as an important plant area (Murphy et al. 2022).
The survival of Microcos rumpi may not seem urgent at present because while data on habitat loss around the Rumpi Hills shows a steady increase (Beckline et al. 2018), there are also still large tracts of habitat seemingly intact where the species may survive (Google Earth imagery, accessed July 2022). However, scenarios can change very rapidly, as with Saxicolella deniseae which when collected for the first time in 2018 was not considered to be at risk of extinction, but which appears to have become globally extinct in 2020 or 2021, before it was published (Cheek et al. 2022a).
Yaoundé and we thank the late Dr Benoît Satabié, Drs Gaston Achoundong, Florence Ngo Ngwe, Eric Nana, Jean Lagarde Betti, the current and former leaders, of IRAD-National Herbarium of Cameroon, Yaoundé, for expediting the collaboration between our two institutes. Dr Duncan Thomas is thanked for collecting the specimens and for providing helpful comments on the Rumpi Hills TIPA datasheet compiled by Dr Bruce Murphy, which gave useful context for the discussion in this paper. Rosemary Lomer, volunteer with the Cameroon TIPAs project, uncovered the specimens of what turned out to be the new species, and brought them to attention: without her this species would still remain unknown to science. Janis Shillito typed the manuscript. Two anonymous reviewers are thanked for constructively reviewing an earlier version of this paper.

Conflict of interest
The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.