Kalanchoe variifolia (Guillaumin & Humbert) Shtein, D.

Shtein, Ronen, Klein, David-Paul & Smith, Gideon F., 2021, A new status for two varieties previously included in the southern Malagasy Kalanchoe rosei, now included in K. variifolia (Crassulaceae subfam. Kalanchooideae), Phytotaxa 496 (3), pp. 228-244 : 233-237

1. Kalanchoe variifolia (Guillaumin & Humbert) Shtein, D. -P.Klein & Gideon F.Sm., stat. nov.

Basionym: Kalanchoe rosei subsp. variifolia Guillaumin & Humbert (1941: 319) .

Homotypic synonym:— Kalanchoe rosei var. variifolia (Guillaumin & Humbert) Shaw (2008: 27) .

Type:— MADAGASCAR. Behara , 1939, Boiteau f. 227 [MNHN-P-P00443995; image available at http://coldb.mnhn.fr/catalognumber/ mnhn/p/p00443995] (lectotype, P!) here designated as lectotype .

Epitype:— MADAGASCAR. Behara, July 1936, Botanical Garden of Tananarive with P. Boiteau no 2027 [MNHN-P-P00443996; image available at http://coldb.mnhn.fr/catalognumber/mnhn/p/p00443996] (epitype, P!) here designated as epitype .

Notes on the lecto- and epitype:—In a standalone paragraph in the protologue of K. rosei subsp. variifolia, Guillaumin & Humbert (1941: 319) cite the following specimen: “ Madagascar Sud: Behara (Boiteau f. 227, 1939)” [English: “Southern Madagascar: Behara (Boiteau f. 227, 1939)”]. In the very next paragraph, they state: “La description cidessus, faite sur le vif, doit être complétée par les observations suivantes effectuées sur l’échantillon d’herbier du Jardin botanique de Tananarive n° 2027 provenant également de Behara: […]”. [English: “The description above, made on the spot [i.e., in the field], must be supplemented by the following observations made on the herbarium specimen from the Botanical Garden of Tananarive no 2027 also from Behara: […]”].

The specimen Boiteau f. 227 was collected in 1939, two years before K. rosei subsp. variifolia was described, but it is not known whether Boiteau, who had a strong interest in Kalanchoe , was present when the description was drawn up “on the spot”. Boiteau was the Director of Parks and Gardens, Tananarive, Madagascar, from 1935 to 1938 and Director of the Parc Botanique et Zoologique de Tsimbazaza (PBZT) from 1935 to 1947 ( Dorr 1997: 47). A further possibility is that Guillaumin and Humbert were aware of the existence of the specimen Boiteau f. 227 and drew up a description when they encountered material in the field that agreed with this Boiteau specimen.

Furthermore, Guillaumin & Humbert (1941: 319) essentially provided two descriptions for K. rosei subsp. variifolia : a Latin one, which was a requirement for the valid publication of plant names at the time ( Turland et al. 2018: Art. 39.1), as well as an amplified description in French. This was not such an unusual step as plants were often briefly or at least in abbreviated form, diagnosed or described in Latin (when that was a requirement; see Turland et al. 2018: Art. 39.2), with longer, more detailed descriptions provided in the native tongue of the author, or the language preferred by the journal in which the name was published. The Latin description of K. rosei subsp. variifolia was based on Boiteau f. 227, while the additional description in French was based on a specimen referred to as Botanical Garden of Tananarive no 2027 ( Guillaumin & Humbert 1941: 319).

Boiteau f. 227, which clearly forms part of the original material, is held at Herb P. under MNHN-P-P00443995 and is available for examination online at http://coldb.mnhn.fr/catalognumber/mnhn/p/p00443995. Two mostly handwritten labels, apart from the barcode sticker, are attached to this specimen. On the label attached to the bottom right corner, it states that the specimen had been taken up in “Hort. Bot. Parisiensis Collection” in “1941” and the specimen is identified as “ Kalanchoe Rosei Hamet subsp. variifolia Humbert & Guillaumin ” [note the sequence of the authors; the subspecies was eventually described by Guillaumin & Humbert, in the opposite order, therefore]. The specimen is further noted as having been collected as “Boiteau f. 227, 1939”. The second label, completely handwritten, perhaps by Pierre L. [“Rapierre”] Boiteau, is pinned to the bottom left of Boiteau f. 227, and states “Les feuilles en sachet proviement de rameaux non fleuris du même individu.” [English: “The leaves in [the plastic] sachet [attached to the specimen] is from non-flowering twigs of the same individual.”].

A further specimen, Botanical Garden of Tananarive no 2027, is also clearly referenced in the protologue of K. rosei var. variifolia and it too constitutes original material. At least two specimens are deposited at Herb. P under Botanical Garden of Tananarive no 2027; both specimens are available for examination online. These specimens are: (1) MNHN-P-P00443996; http://coldb.mnhn.fr/catalognumber/mnhn/p/p00443996; and (2) MNHN-P-P00443997; http:// coldb.mnhn.fr/catalognumber/mnhn/p/p00443997. In the online catalogue of Herb. P., these specimens, barcoded as MNHN-P-P00443996 and MNHN-P-P00443997, respectively, are recorded as “Jardin Botanique Tananarive|Boiteau, P. With P. Boiteau. 2027”. To prevent confusion between the two specimens, Boiteau f. 227 and Botanical Garden of Tananarive no 2027, we here do not refer to Botanical Garden of Tananarive no 2027 as “ Boiteau 2027 ”. There is some evidence that corroborates that Boiteau indeed collected Botanical Garden of Tananarive no 2027 (see discussion on photographs, below).

Labels that give the collecting date as “juil. 36” and “ Juillet 1936 ” are attached to MNHN-P-P00443996 and MNHN-P-P00443997, respectively. Both specimens were collected from “Behara”, southern Madagascar, without further information, and with the same coordinates [of Behara, i.e., 24°57’0’’ S 46°22’59.988’’ E] catalogued for them in the online version of Herb. P. It is likely that these coordinates were only later recorded for the specimens, based on the narrative, “Behara”, as mentioned on the specimen labels. One of these specimens, MNHN-P-P00443996, is provided with the identification “ Kalanchoe Rosei Hamet et Perr. subsp. variifolia Guill. et Humb. ”, therefore with the authors listed in the sequence in which the variety was eventually described. The other specimen, MNHN-P- P00443997, was only identified as “ Kalanchoe ”. Labels stating “ TYPE ” in bold, red lettering are attached to both specimens .

Two photographs of two virtually identical living plants—the one in flower, the other not—are associated with MNHN-P-P00443996. These two images are clearly labeled as “Photo du Type ” of K. rosei subsp. variifolia , and one of them directly mentions “ P. Boiteau 2027 ” followed by a location and a year, “Behara, 1939”. The date, 1939, most likely corresponds to the year during which the material was collected and the material depicted does not match material preserved as Botanical Garden of Tananarive no 2027, but rather coincides with that of Boiteau f. 227 (see above). From the examination of this second photograph and Boiteau f. 227, especially of a few uniquely deformed leaf apexes and unusual stem dents, it is clear that Boiteau f. 227 was harvested from the plant depicted in that photograph ( Fig. 5 ). We therefore conclude that the two photographs currently associated with Botanical Garden of Tananarive no 2027 instead should be associated with Boiteau f. 227, a different gathering, as mentioned above. The photographs were taken in “ avril 1941 ”, one month before the description of K. rosei var. variifolia was published, and are of material cultivated at the “Muséum”, i.e., the Muséum national d’Histoire naturelle. Although the photographs were therefore taken two years after Boiteau f. 227 was collected, they qualify as original material that was available to Guillaumin & Humbert (1941: 319), which was published in May 1941.

From this analysis it is clear that, even although Guillaumin & Humbert (1941: 319) might have intended Boiteau f. 227 as the type of K. rosei subsp. variifolia , they did not make this clear and two syntypes, Boiteau f. 227 and Botanical Garden of Tananarive no 2027, are in existence, with at least one duplicate also being available of the latter, i.e., an isosyntype. We designate the specimen Boiteau f. 227 as the lectotype of K. rosei subsp. variifolia .

A specimen of Botanical Garden of Tananarive no 2027 [MNHN-P-P00443996] is designated as epitype.

Included taxa:—Two. K. variifolia var. variifolia and K. variifolia var. seyrigii .

Diagnosis:—Medium-sized, perennial, induced-phyllo-bulbiliferous Kalanchoe distinguishable from all other such species by a combination of traits: leaves subsessile to petiolate; leaf blade ± flat, ovate to almost linear, generally at least 2.5× longer than wide; base attenuate to truncate, rarely auriculate; margins strongly dentate-serrate; apex obtuse to acute; calyx papery, 2.5–3× longer than wide; calyx tube at least twice as long as free sepal segments, subcylindrical-urceolate; corolla 25–35 mm long, about twice as long as the calyx, orange to red, sometimes somewhat pink, stipitate, barely constricted at base of petals; stipe wide and short.

Description:— Plants perennial, entirely glabrous, sometimes slightly glaucescent throughout, often unbranched, sometimes basally branched, erect, often basally decumbent, about 0.5–1.0 m tall. Stem 4.0– 7.5 mm in diam., 3–5 mm towards inflorescence, green to brown or purple, cylindrical, suffrutescent to barked in lower part, often developing aerial roots; leaf scars rounded-crescent, conspicuous, slightly protruding, not connate; internodes short in lower ½, 1–3 cm long. Leaves opposite, decussate, sessile to distinctly petiolate, succulent, uniformly green to brown or bluish green, or in green-to-red gradient; petiole varying in length and cross-section dimensions during different phenological stages, not widened basally, subcylindrical to flattened, short and indistinct to long and distinct (up to ± 2 cm long), gradually widening towards blade to slender and isodiametric throughout; blade (3.0–)4.0–7.5(–10.0) × 0.5–4.5 cm, varying in shape during different phenological stages, typically widest in basal ½ or medially, obovateovate or oblanceolate to oblong-elliptic or lanceolate-linear, ± flat, often slightly concave above and convex below, not guttered, abaxial surface not especially coloured, unmarked; margins red to purple, irregularly serrate-dentate or serrulate-denticulate for at least apical ½ or entirely, blotched dark purple between serrations, within blotches marked by pale organogenic and/or embryogenic notches prior to bulbil production or bulbils proper; serrations 2–6 × 3–10 mm, highly variable, increasing in width apically and acuity basally, smaller secondary serrations sometimes present [in K. variifolia var. seyrigii ], long and deep, few or many, sparse or dense, broad or fine, obtuse to very acute; bulbil production induced, possible at each marginal leaf notch, but most common in apical ½; base gently attenuate to ± truncate, entire to minutely auriculate-dentate; apex rounded or abruptly obtuse to acute. Inflorescence dichotomously branched corymbiform cyme, not leafed at anthesis, non-bulbiliferous i.e., does not form dense bulbil clusters, but may form few , large, discrete plantlets, dense, terminal, 5–15 cm wide; peduncle single, cylindrical, 10–25 cm long, 3–5 mm in diameter, nodes few and long; bracts small, ± 3.0 × ± 0.5 cm, sessile, attenuate, linear-lanceolate, subentire or few-dentate apically, otherwise leaf-like, wilting soon, floral bracts smaller, sooner-wilting; pedicels 7–15 mm long, curved, cylindrical, tapering towards flowers from 0.8–1.0 mm to about 0.4–0.6 mm in diam. Flowers many, 27–35 mm long, pendent, campanulate. Calyx 13–18 × 4–6 mm, uniformly green or in gradient of green to brown or bluish purple, especially free sepal segments, thin, papery, subcylindrical-urceolate, rarely slightly 4-angled, basally ± flat to rounded, widest medially or in apical ½, slightly constricted at base; tube 9–12 mm long, nearly isodiametric; free sepal segments narrowly deltoid, apiculate, 4–6 mm long, widest at bottom at 3–4 mm, straight or spreading. Corolla 26–33 mm long, golden-yellow to orange-red or pink-red, usually less vibrant towards calyx, green where obscured by calyx, somewhat stipitate, stipe ± 1 × 2 mm, short, wide; tube 20–28 mm long, ± 4-angled especially basally, basally adpressed around and above carpels up to 5 mm wide, then diminishing to 3 mm wide to its thinnest, then gradually inflating to base of petals where it is very slightly constricted, widest in apical ⅓ or at base of petals at 5–6 mm width; lobes 6–8 × 4–5 [at base] mm, widest at 5–6 mm at ⅓ from tip, oblong-obovate, apically rounded, very slightly apiculate, usually somewhat spreading. Stamens 8, inserted low down in corolla tube, at ± level of carpels, included, barely reaching lower ⅓ of petals; filaments of two similar lengths, adpressed, protruding inside base of corolla, free for ± 17–21 mm, thin, dull green basally and apically, strongly infused with orange-pink later yellow; anthers ± 1.0 × 0.7 mm, dark grey, narrowly ovate-cordate, reaching lower ⅓ of petals. Pistil consisting of 4 carpels; carpels ± 6.0–7.5 × 2.0– 2.5 mm, convergent, fused basally, free above, uniformly dark green, ovate-oblong, widest in lower ⅓, attenuate towards styles; styles ± 12–18 mm long, uniformly yellowish-green; stigmas very minutely capitate, dark green; scales ± 1.7–2.5 × 1.0– 1.5 mm, brown- or purple-green, free, lanceolate-narrowly trapezoidal, longer than wide, thickened basally, tapering apically in thickness and width, incurved, upper corners obtusely truncate. Follicles not seen. Seeds not seen. Chromosome number: 2 n = unknown.

Notes on description:—The description of K. variifolia is primarily based on the nine specimens that constitute the original material of the two varieties recognised, i.e, the autonymic one and K. variifolia var. seyrigii , as well as their protologues, observations made on other specimens held at Herbs P and B, and several living plants. Note though that the description included in the protologue of K. rosei var. variifolia , the basionym of K. variifolia , does not provide much detail, and for the most part emphasises how it can be distinguished from the earlier-described K. variifolia var. seyrigii . Usefully, much of the description could be verified against the more detailed description of the latter variety. Guillaumin & Humbert (1941: 319) did not record any differences between the flowers of the two varieties, which agrees with our observations of the preserved and living material included in this study (see ‘Traits insufficient to distinguish the varieties’ below for more detail).

Key to the varieties of Kalanchoe variifolia :—In the following key, traits only observable in living material are denoted by an asterisk (*). Other characteristics are based on observations of specimens and the protologues of the names (see above).

1a. Plants not glaucescent*; stem early-deciduous*, basally especially woody, bark present; leaf apex abruptly obtuse, terminating in a blunt tooth; margins sparsely serrate-dentate, for at least the apical ½ to fully so, serrations or dentations rarely doubly superimposed; calyx turning papery after the corolla wilts*..................................................................... K. variifolia var. variifolia

1b. Plants slightly glaucescent*; stem late-deciduous*, basally somewhat woody, bark lacking; leaf apex abruptly acute or rounded, terminating in a sharp tooth; margins densely serrulate-denticulate throughout, serrulations or denticulations often doubly superimposed; calyx turning papery before corolla wilts*........................................................................... K. variifolia var. seyrigii

1a. K. variifolia ( Guillaumin & Humbert 1941: 319) Shtein, D. -P.Klein & Gideon F.Sm. var. variifolia

Type:—same as for Kalanchoe rosei subsp. variifolia , basionym of K. variifolia ( Turland et al. 2018: Art. 7.7).

Description of K. variifolia var. variifolia :— Plants not glaucescent; stem early-deciduous, especially basally woody, barked; petiole generally 5–20 × 2.5–5.0 mm if present, greatly widening apically; leaf blade 4–10 × 1.2–4.0 cm, varying from narrowly obovate-oblanceolate to narrowly elliptic to broadly lanceolate through phenological phases; base attenuate to cuneate; apex abruptly obtuse, terminating in a blunt tooth; margins serrate-dentate, for at least apical ½ to fully so, serrations or dentations rarely doubly superimposed; serrations 5–10 × 3–6 mm, few, especially long and deep, broad, sparse, subacute to obtuse. Calyx usually green or brown, turning papery after corolla wilts.

Additional material examined: —P00443999, P00444004, P00444005 [Muséum national d’Histoire naturelle, Paris ( France), Collection: Vascular plants (Herb. P)]. B 10 1186338, B 10 1186346, B 10 1186347 [Herbarium Berolinense (B), Botanischer Garten und Botanisches Museum Berlin-Dahlem, Berlin].

Distribution: —Forested areas near Behara, southeastern Tuléar, southern Madagascar.

1b. K. variifolia var. seyrigii (Boiteau & Mannoni) Shtein, D. -P.Klein & Gideon F.Sm., comb. nov.

Basionym:— Kalanchoe rosei var. seyrigii Boiteau & Mannoni (1949: 71) , as ‘ Seyrigi ’.

Homotypic synonyms:— Kalanchoe rosei subsp. serratifolia Humbert (1933: 166) . Kalanchoe rosei var. serratifolia (Humbert) Shaw (2008: 27) , nom. illeg. ( Turland et al. 2018: Art. 11.2).

Type:— MADAGASCAR. Toliara, 4–5 km north of Fort-Dauphin , towards the lagoon, max. 25 m asl, 20 September–6 October 1928, H. Humbert 5979 [P P00374204; image available at http://coldb.mnhn.fr/catalognumber/mnhn/p/p00374204] (lecto., P!) here designated as lectotype .

Nomenclatural note:—At the rank of variety, K. variifolia var. seyrigii , not ‘ K. variifolia var. serratifolia ’, should be adopted because a name, in this case K. rosei subsp. serratifolia , does not have priority outside the rank at which it was published ( Turland et al. 2018: Art. 11.2).

Note on the type:—In the protologue of K. rosei var. seyrigii, Humbert (1933: 166) cited a single gathering, H. Humbert 5979, that was collected during the period 20 September to 6 October 1928 in the vicinity of Fort- Dauphin , southern Madagascar. At least three specimens are deposited at Herb. P under H. Humbert 5979, and another such specimen is held by Herb. B. These are: (1) MNHN-P-P00374204; http://coldb.mnhn.fr/catalognumber/mnhn/ p/p00374204; (2) MNHN-P-P00374205; http://coldb.mnhn.fr/catalognumber/mnhn/p/p00374205; (3) MNHN-P- P00374206; http://coldb.mnhn.fr/catalognumber/mnhn/p/p00374206; and (4) B 10 0153778; http://herbarium.bgbm. org/object/B100153778. Images of all four specimens are available for examination online, with the Herb. B specimen being the most depauperate. All the specimens carry virtually identical handwritten labels. The word “ Type ” is included on all four labels .

We interpret these specimens as a single gathering of a single taxon, K. variifolia var. seyrigii , prepared at the same time from a single locality. All these specimens are syntypes. We here designate MNHN-P-P00374204 held at Herb. P as lectotype. Note that the coordinates shown on the MNHN website for this gathering are erroneous, and do not match the location of the type locality provided in the protologue of K. variifolia var. seyrigii or on the labels attached to each of the specimens, but instead point to a place almost 130 km west of Tolagnaro.

Later, Guillaumin & Humbert (1941: 319) noted that parts of this material, Humbert 5979, were cultivated in Antananarivo, under Jardin Botanique Tananarive 2029. We assume it was derived clonally, but the material preserved as Jardin Botanique Tananarive 2029 was only collected in July 1936, some three years after K. rosei var. seyrigii was described. At least two specimens are deposited at Herb. P under Jardin Botanique Tananarive 2029: (1) MNHN-P- P00444000; http://coldb.mnhn.fr/catalognumber/mnhn/p/p00444000; (2) MNHN-P-P00444001; http://coldb.mnhn.fr/ catalognumber/mnhn/p/p00444001. An image of one of the two, P-P00444001, is available for examination online .

Description of K. variifolia var. seyrigii :— Plants slightly glaucescent; stem late-deciduous, basally suffrutescent, not barked; petiole generally 15–20 × 2.5–3.5 mm if present, subcylindrical throughout; leaf blade 3–5 × 0.5–1.5 cm, varying from ovate-lanceolate or obovate-oblong to lanceolate to linear or elliptic throughout phenological phases; base cuneate to truncate; apex abruptly acute or rounded, terminating in a sharp tooth; margins serrulatedenticulate throughout, serrations or dentations often doubly superimposed; serrations 3–5 × 2–3 mm, numerous, deep, comparatively fine, often dense or sparse but double, very acute. Calyx usually tanned brown or purple, turning papery before corolla wilts.

Additional material examined: —B 10 1068977, B 10 0148415, B 10 1186343, B 10 1186344, B 10 1186345 [Herbarium Berolinense (B), Botanischer Garten und Botanisches Museum Berlin-Dahlem, Berlin].

Distribution: —Forested areas near Tolagnaro and the coastal area in southeastern Tuléar, southern Madagascar.

Figures

FIGURE 5. Comparison of the available original material of Kalanchoe variifolia var. variifolia at Herb. P. The scale of A–B is based on photograph size; B–D to scale. A–B. The two photographs held together with Jardin Botanique Tananarive et P. Boiteau 2027 (P- P00443996, P-P00443997). C. Boiteau f. 227 (P-P00443995), the lectotype here designated for K. variifolia var. variifolia, which is clearly the exact same plant as figured in element B. D. The sessile leaves and lower peduncle of Jardin Botanique Tananarive et P. Boiteau 2027 (P-P00443996), a plant separately collected. Photographs of material held by Herb. P (Vascular plants collection, Muséum national d’Histoire naturelle, Paris).