﻿Taxonomy and nomenclature of Abutilonalbidum (Malvaceae, Malvoideae), a cryptic Saharo-Canarian species recently rediscovered in Tenerife

﻿Abstract Abutilonalbidum, a cryptic Saharo-Canarian species, was thought to have been last collected in 1945 in Tenerife by E.R. Sventenius. In 2019, it was rediscovered in the same area. The characteristic features of the Canarian plants are discussed, especially in relation to the morphologically similar-looking and probably closely-related species Abutilonindicum and A.bidentatum. It is concluded that the plants from Tenerife and north-western Africa indeed represent a distinct species. The species is illustrated and a key for the identification of this and related species is presented.


Introduction
Abutilon Mill. (Malvaceae, Malvoideae) is a large genus which mostly occurs in tropical and subtropical areas around the world, with a few species extending to warm-temperate regions. The genus is most diverse in the Neotropics (Areces Berazaín and Fryxell 2007). It is one of the largest Malveae genera, although the number of its recognised species considerably varies, depending on the species delimitation. Most authors nowadays accept ca. 150-160 species (e.g. Bayer and Kubitzki (2003); Mabberley (2008); Verdcourt and Mwachala (2009)), but the boundaries of the genus remain unclear.
In their monumental "Histoire Naturelle des îles Canaries", Webb and Berthelot (1836) provided a detailed description of "Abutilon albidum Nob.", a species they explicitly based on Willdenow's Sida albida Willd. It was reported from the Santa Cruz area in the south-eastern corner of the island of Tenerife. Subsequent authors (e.g. Bornmüller (1904); Pitard and Proust (1908); Lindinger (1926); Burchard (1929)) regularly confirmed its presence in a small area roughly located between Santa Cruz and Igueste de San Andrés. However, after the 1940s, it was no longer collected nor mentioned in the regional literature. Surprisingly and apparently without any explanation, Canarian records of this species were later considered to be referrable to the invasive weed A. grandifolium G. Don (e.g. Eriksson et al. (1974); Santos (1983); Hohenester and Welss (1993); Valdés (2011); Dobignard and Chatelain (2012); BIOTA (2023)).
In December 2019, an unknown species of Abutilon was observed by one of us (F.V.) in several localities in the village of Igueste de San Andrés (Municipality of Santa Cruz de Tenerife), on the verge of the natural protected area 'Anaga Rural Park' in the south-easternmost part of Tenerife. The observed plants clearly differed from another species of Abutilon that is widely naturalised and invasive in the Canary Islands, A. grandifolium. Although identification attempts were at first unsatisfactory, the plants eventually were determined to belong to the species that had been known from Tenerife since the end of the 18 th century, but which had not been collected since 1945 and apparently completely forgotten by Canarian authors.

Materials and methods
Fieldwork was undertaken by the first author in December 2019 and March 2022 and, for revision of other localities in Tenerife, by the second author in May and June 2021. Herbarium specimens were collected and deposited at the herbarium of the Meise Botanic Garden [BR; for herbarium acronyms Thiers (2023) is followed]. In addition, numerous photographs were made that showed essential diagnostic features, which are less easily observed in pressed specimens. In March 2022, a detailed inventory was made of all remaining populations and all were georeferenced and registered in the nature observations platform Observation.org (https://observation.org/soort/ view/1029740).
Herbarium specimens relevant for this study (including type material) were studied in the herbaria of the Meise Botanic Garden (BR), the Jardín Botánico Canario Viera y Clavijo (LPA) and the Jardín de Aclimatación de la Orotava (Instituto Canario de Investigaciones Agrarias) (ORT). In addition, images from several herbaria available online were also consulted (B, COI, FI, G, K, L, MO, MPU, NYBG, P, RAB, UM and W).
Countless literature sources (including protologues), deemed useful for this study, were checked to better understand the identity and taxonomy of the species and the characters that differentiate it from closely-related taxa.

Results
The nomenclature of Abutilon albidum Willdenow (1809) validly published the name Sida albida Willd. for a species that was later transferred to Abutilon by Sweet (1826) as A. albidum (Willd.) Sweet. The only original extant specimen for S. albida is seemingly the holotype (Fryxell 2002) and is preserved at B. From this specimen and the protologue, it can be deduced that the plant described as S. albida definitely differs from the Saharo-Canarian one.
However, since there is no other binomial for the Saharo-Canarian plant, conservation of the name Sida albida with a conserved type was proposed (Verloove and Sennikov (2022) and discussion therein). If the proposal is not accepted, a new name will have to be given to this species.
The quest for the identity of an unknown Abutilon from Tenerife From Tenerife, only a single, yellow-flowered species of Abutilon is currently known, A. grandifolium. This species, native to South America, is widely naturalised and invasive in the Canary Islands. Plants found in 2019 in the south-eastern part of Tenerife are strikingly different. In comparison to A. grandifolium, their corollas are much smaller, fruit (schizocarp) has more numerous mericarps (usually around 15), calyx is much shorter than the fruit, stem indumentum is quite different (a mixture of short stellate hairs and long patent hairs) and inflorescences consist of both solitary flowers and axillary (pseudo-)panicles. Yet, since the second half of the 20 th century, Canarian authors considered historical records of A. albidum to be erroneous and referred them to A. grandifolium instead (Eriksson et al. 1974;Santos 1983;Hohenester and Welss 1993;Valdés 2011;Dobignard and Chatelain 2012;BIOTA 2023).
If the Tenerife plant is not Abutilon grandifolium, then what is its correct name? Identification attempts initially led -with not too much difficulty -to A. indicum, a common weed from the Old World tropics (e.g. Borssum Waalkes (1966)) that is also widely naturalised in the Caribbean (Fryxell 1989). However, the Tenerife plant clearly differed from that species as well, for instance, in having much shorter petals and mericarps. Its stem indumentum is also quite different: it is composed of a mixture of very short, dense stellate hairs, sparser and slightly longer multicellular hairs and rather numerous long, patent simple hairs, whereas in A. indicum the stem is densely covered with stellate hairs, rarely with some additional slender, simple hairs.
Identification keys from Pakistan and India key out Abutilon indicum together with a similar species, A. bidentatum (Hochst.) A. Rich., described from Ethiopia, but widely distributed from Central Africa to India. In areas where these two species occur sympatrically, they are considered to be very similar. Husain and Baquar (1974) separated A. bidentatum from A. indicum, based on its smaller, paler corollas. Bhandari (1995) distinguished A. bidentatum and A. indicum, based on carpel length (6-8 mm vs. 9-12 mm, respectively). According to Paul and Nayar (1988), both species are differentiated as follows: A. bidentatum has a staminal column 2-3 mm long, a schizocarp 10 mm across, mericarps 10 × 5 mm and gradually acuminate dorsally, whereas A. indicum has a staminal column 5-7 mm long, a schizocarp 1.5-2.5 cm across, mericarps 10-15 × 7-10 mm and acute to acuminate dorsally. The diagnostic characters put forward by these authors for A. bidentatum are evidently more in line with those observed in the plants found in Tenerife. However, while A. bidentatum has mericarps with two distinct dorsal cusps (hence the species epithet), the mericarps in the Tenerife plants are merely shortly acute-triangular dorsally at maturity, without clear protuberances.
Eventually, floristic accounts from north-western Africa, more precisely from Algeria and Morocco (Quézel and Santa 1963;Fennane et al. 1999), which are geographically close to Tenerife, threw new light on the possible identity of the Tenerife plant. Based on mericarp number and stem indumentum, the plants were easily keyed out as Abutilon albidum subsp. albidum. A subsequent study of herbarium specimens (including type material) and protologues of Sida albida (Willdenow 1809), S. bidentata (Hochstetter 1842) and S. indica (Linnaeus 1756) demonstrated that these three spe-cies are very similar in many characters. Yet, the plant material from Tenerife and north-western Africa is considered to be sufficiently distinct, as correctly assumed by Webb and Berthelot (1836) and other Canarian workers up to the first half of the 20 th century (e.g. Bornmüller (1904); Pitard and Proust (1908); Lindinger (1926); Burchard (1929)) and by contemporary North African authors (Fennane et al. 1999). Preliminary results from a molecular analysis also showed these three species to be closely related, but distinct (Verloove, unpubl. data).

Diagnostic characters
Morphological features used for the separation of Abutilon albidum, A. bidentatum and A. indicum are discussed below.
The stem indumentum of these species is different. Abutilon indicum has an indumentum that almost entirely or even exclusively consists of very short and dense stellate hairs. Simple, long patent hairs are always absent or very sparse (e.g. Hochreutiner (1902); Adams (1972); Townsend (1980);Fryxell (1989)). Abutilon albidum and A. bidentatum, on the contrary, have a mixture of short stellate hairs and long patent hairs; both hair types are equally abundant. In addition, rather numerous multicellular hairs are observed (some of them gland-tipped). In length, these hairs are slightly or much longer than the stellate ones. This feature is rarely mentioned in floristic accounts, although exactly the same pubescence is observable in the syntype of A. bidentatum kept at BR. Notably, Hochreutiner annotated a syntype at K in 1899: "Le A. bidentatum de Hochst. est une forme villeuse et légèrement glanduleuse de A. indicum Sw. avec lequel nous l'assimilons" [translation F.V.: A. bidentatum of Hochst. is a villous and slightly glandular form of A. indicum Sw. with which we equate it.]. Apparently, Hochreutiner also noticed the presence of the third hair type in A. bidentatum. Wood (1997) described A. bidentatum as a glandular perennial and this likely also refers to the presence of multicellular hairs. The long description of Webb and Berthelot (1836) did not refer to the presence of multicellular hairs. However, such hairs are observable in historical, as well as recent, collections from Tenerife.
These three species also differ in flower size and colour. The populations found in Tenerife are small-flowered: flowers are ca. 15 mm in diameter with petals 7-10 mm long (i.e. only slightly longer than the calyx). Abutilon indicum, in contrast, is a largeflowered species with corollas 25-35 cm in diameter and with petals 12-15 mm long (e.g. Borssum Waalkes (1966); Abedin (1979); Naqshi et al. (1988); Philcox (1997); Areces Berazaín and Fryxell (2007); Bano and Deora (2018)). Petals are, in fact, about twice as long as the calyx or even longer (Riedl 1976). In the syntype of A. bidentatum at BR, flowers are about 18 mm in diameter. In addition, the staminal column is much shorter in A. albidum and A. bidentatum than in A. indicum (e.g. Abedin (1979); Naqshi et al. (1988); Paul and Nayar (1988)). The plants from Tenerife and a syntype of A. bidentatum at BR have a staminal column 2-3 mm long, whereas the staminal column in A. indicum is 5-7 mm long. Finally, petals of A. albidum and A. bidentatum tend to be slightly paler; i.e. pale yellow to yellow ('pallidè luteola' for A. albidum, according to Webb and Berthelot (1836)), whereas those of A. indicum are usually said to be yellow to orange yellow (e.g. Husain and Baquar (1974); Abedin (1979); Naqshi et al. (1988)).
Fruit characters are also different in these species. Mericarps are invariably longer and wider in Abutilon indicum. In areas where this species and A. bidentatum occur sympatrically they are distinguished (often even exclusively) based on this character (e.g. Jafri (1966); Paul and Nayar (1988); Bhandari (1995)). In the plants from Tenerife, mericarps are up to 9 mm long and 5 mm wide (often slightly smaller). This is roughly in line with measurements taken on the syntype of A. bidentatum at BR (8 mm long and 5 mm wide). In A. indicum, in contrast, mericarps are often twice as long and broader. The number of mericarps per fruit tends to be higher in A. indicum than in A. albidum and A. bidentatum. In A. albidum, the number of mericarps ranges between 10 and 15 (Webb and Berthelot 1836), although most fruits have around 15 mericarps. Similar numbers are known for A. bidentatum, whereas Philcox (1997) gives 20-30 for A. indicum. We observed some overlap in these measurements, but the mericarps are, indeed, more numerous in A. indicum. In all these species, mericarps turn black at maturity; they are hairy outside and shiny inside. Mericarp ornamentation is considered to be an important diagnostic feature for the separation of species in Abutilon. The outer apical (dorsal) margin can be either rounded (without protuberances) or angled. In the latter case, the outgrowth can be merely acute, gradually acuminate to aristate. Abutilon bidentatum was initially described as having mericarps that are bidentate at apex ("carpellis compressis apice truncatis, bidentatis"). In the syntype at BR, a cusp ca. 1 mm long is discernible. Most authors describe the cusp length as 1-2 mm long (e.g. Thulin (1999); Verdcourt and Mwachala (2009)). At this point, the plants observed in Tenerife definitely differ from A. bidentatum: mericarps are angled at the dorsal margins, but a cusp or protuberance is missing. In A. indicum, mericarps can vary from rounded, obtuse to long acuminate (Borssum Waalkes (1966); see also Bano and Deora (2018)). The schizocarp also tends to be less wide in A. albidum and A. bidentatum (10-15 mm across) than in A. indicum (rather 15-25 mm across). Finally, according to Jafri (1966), mericarps in A. bidentatum already dehisce before breaking away from the central axis -a feature also observed in the plants from Tenerife -whereas in A. indicum mericarps dehisce after breaking away from the central axis.
Finally, Abutilon albidum, A. bidentatum and A. indicum also differ in inflorescence shape. All have flowers that are inserted in the leaf axils. However, whereas in A. indicum flowers are invariably solitary, they often merge into distinct panicles in A. bidentatum (e.g. Verdcourt and Mwachala (2009)). This feature was also observed in the plants that are found in Tenerife.
On the other hand, some diagnostic characteristics mentioned in literature proved to be of no value or merely erroneous. For instance, pedicels are usually said to be shorter than petioles in A. bidentatum and vice versa in A. indicum (Abedin 1979;Borssum Waalkes 1966;Ghazanfar 2003;Verdcourt and Mwachala 2009). In reality, pedicel and petiole length are variable in these three species, although pedicels often are at least as long as or longer than petioles, also in A. bidentatum (as can be seen in a syntype at BR, as well as in the material from Tenerife).
The features that appear to be most reliable for the separation of these three species are summarised in Table 1.
From the above, it can be concluded that Abutilon albidum is most closely similar to A. bidentatum morphologically. These two species share all characters, except that their mericarps are ornamented in different ways. Since mericarp ornamentation is considered to be an important diagnostic trait in the genus (Alzahrani et al. 2021), both species are apparently distinct. Besides these morphological differences, the species also seem to differ ecologically and occur in non-overlapping areas (see below).
Finally, another species that has often been associated with Abutilon albidum is A. fruticosum Guill. & Perr. In fact, Maire (1933) reduced the latter to subspecies rank, as A. albidum subsp. fruticosum (Guill. & Perr.) Maire (see also Quézel and Santa (1963)). A specimen of A. albidum at FI, annotated with 'type', is also stored under A. fruticosum. More or less similar criteria are also followed by POWO (2023) that considers A. albidum sensu Webb & Berthel. as a synonym of A. fruticosum. Hochreutiner (1902) already emphasised that these two species are, in fact, not related. Abutilon fruticosum has smaller and paler (green turning brown, never blackish), almost turbinate fruits that always have fewer mericarps (usually 9-10), smaller rounded leaves, often with almost entire margins and a stem indumentum that entirely consists of very short stellate hairs.  Abedin (1979); Vollesen (1995); Ghazanfar (2003); Verdcourt and Mwachala (2009);Randall (2017) and many other sources). Flowers 25 Description. Erect, short-lived perennial herb or shrub up to 60(-100) cm tall. Branches densely covered by minute stellate hairs, mixed with numerous long, simple spreading hairs (especially, but not exclusively on new growth) and sparser multicellular hairs. Leaves (median cauline) up to 13 cm long and 7 cm wide, broadly ovate to almost rotund, deeply cordate at base, acute to acuminate at apex, irregularly crenatetoothed to double-toothed, very shortly stellate-pubescent on both sides; petiole up to 6 cm long, slightly shorter than blade, very densely stellate hairy mixed with rather numerous short, glandular, capitate hairs and scattered weak, simple, spreading hairs; longest stipules to 9 mm long (mostly shorter), filiform. Flowers axillary, mostly solitary, but sometimes merging to more or less distinct panicles, pedicel 2-4 cm long, elongating in fruit up to 1-8 cm, articulate and geniculate ca. 5 mm below the apex. Calyx 5-lobed for ca. ½ its length, sepals up to 8 mm long, densely long-pubescent on both sides, slightly accrescent, erect, ultimately reflexed; lobes lanceolate-acuminate. Corolla pale yellow to yellow, ca. 1.5 cm across; petals 7-10 mm long, 4-5 mm wide, obovate. Staminal column 2-3 mm long, stellate-pubescent. Fruit 8-10 mm long, ± 13 mm across; mericarps 10-15, black and rather star-like spreading at maturity, up to 9 mm long, 5 mm wide, the outer apical (dorsal) angle acute-triangular, without protuberances, stellate-pubescent towards the margin. Seeds 2(-3), brown, initially with scattered stellate hairs, pilose near margins, ca. 2 mm across.  Illustrations. Webb and Berthelot (1836) presented the first superb line drawing of Abutilon albidum. Figs 1-4 show the species in nature in the Canary Islands (Tenerife) and Morocco (Tiglit), respectively.

Character / species
Distribution. Macaronesia (Spain: Canary Islands), north-western Africa (Algeria: Hoggar Mountains; Morocco: Anti-Atlas) (Fig. 6). Webb and Berthelot (1836) reported Abutilon albidum from 'Barranco Santo', near the Zurita Bridge, in Santa Cruz de Tenerife, in the south-eastern part of the Island. It had previously been collected from Tenerife in 1796 by A. de Jussieu (P!; see specimens examined). It was considered a Canarian endemic (Masferrer y Arquimbau 1880). Nearly all collections that have been made since then originated from the Barranco de Santos Ravine (present-day name) in Santa Cruz (Fig. 7). Historical literature sources mostly also referred to that locality (e.g. Masferrer y Arquimbau (1880); Bornmüller (1904)). Some further historical literature sources reported it from additional localities in Tenerife. According to Lindinger (1926), it was also observed in La Orotava, ca. 35 km further SW. Pitard and Proust (1908) also mentioned it from another ravine in Santa Cruz, the Barranco del Bufadero. The species was said to be very rare in Tenerife. Burchard (1929) probably presented the most extensive (and most recent) update of its occurrence on the Island. It was mostly seen between San Andrés and Igueste [F.V.: i.e. Igueste de San Andrés], including below the Los Órganos cliffs [F.V.: i.e. north of the Las Teresitas Beach]. It was also said to have isolated occurrences in Santa Cruz, where it had probably become extinct. All documented occurrences, except for a single record from La Orotava, were from the extreme south-eastern part of the Island where it occupied an area of hardly more than 12 km 2 . In Tenerife, it was probably last collected in 1945 by E.R. Sventenius. According to some sources, it was formerly also collected in Gran Canaria (Barranco de la Angostura) (Pitard and Proust 1908); unfortunately, no specimens are present in the LPA herbarium.
In December 2019, Abutilon albidum was rediscovered in Tenerife in Igueste de San Andrés, 74 years after Sventenius' collection from 1945. In the same village, several small populations of identical plants were observed in December 2019. Some plants were seen along the side of the road Carretera de Igueste de San Andrés adjacent to the Barranco de Igueste (these plants are also visible on Google Streetview images from June 2012: https://www.google.com/maps/@28.5291262,-16.1568142,3a,75y,15.61h,60.37t/ Figure 5. Abutilon grandifolium in Telde, Gran Canaria. April 2017, F. Verloove. For many decades, A. albidum was erroneously considered to be a synonym of this invasive species from South America. Abutilon grandifolium has much larger, more orange-yellow corollas and sepals forming longitudinal keels in bud and saccate at base. data=!3m6!1e1!3m4!1s6-0-oKg77q5kAJSCf1_wCQ!2e0!7i13312!8i6656). Further plants were seen as weeds in a plantation near the San Pedro Apóstol church. Finally, the species was also observed on both sides of the Paseo el Cementerio, a track that  leads to the cemetery. In March 2022, a more detailed survey was carried out (Fig. 8): the species was found in several additional localities, but less than 100 individuals in total and in an area of at most 35,000 m 2 (for details see: https://observation.org/soort/ view/1029740). Interestingly, in the very same area, the Belgian amateur botanist Leon Delvosalle  collected this species in May 1962, but erroneously identified it as Sida rhombifolia, a subtropical Malvaceae weed that is also naturalised there. After the Meise Botanic Garden (BR) acquired his herbarium, this collection (consisting of four small fragments only) was re-identified by the first author as Abutilon cf. indicum in March 2013. All these localities fall within the distribution area as described by Burchard (1929). From this, it can be deduced that A. albidum has survived at least in Igueste de San Andrés in Tenerife. Meanwhile, Santa Cruz de Tenerife became the capital of the Island and has dramatically changed in the past century. Repeated botanical explorations over the years in the ravines where A. albidum formerly occurred, especially Barranco de Santos and Barranco del Bufadero, were fruitless. Additional fieldworks in May and June 2021 in potentially suitable areas in Barranco Tahodio, Barranco Valleseco, between Igueste de San Andrés and San Andrés, Playa de las Teresitas (Bajo Los Órganos) and in Playa de Las Gaviotas (Bajo Los Órganos) were also unsuccessful. The coastal area between Santa Cruz and San Andrés was mostly lost due to the expansion of the commercial port and the urbanisation of the capital. Moreover, the highly invasive grass species Cenchrus setaceus (Forssk.) Morrone has now colonised the entire area, including the habitats where A. albidum once occurred. In Los Órganos, the species probably grew near the old path between Las Teresitas and Las Gaviotas that goes through the base of the basaltic rocks of Los Órganos.
After its discovery in Tenerife, Abutilon albidum also was found in north-western Africa. Maire (1933) first reported it from Ideles and Imarera in the Hoggar Mountains, a highland region in the central Sahara in southern Algeria, along the Tropic of Cancer, where it was said to be rare (Quézel and Santa 1963). Its presence there has not been confirmed lately and it is likely extinct (pers. comm., Prof. Said Amrani, January 2022). From Morocco, it was first reported from Assa in the southwest of the country