﻿A note on the type of Harpalyce (Fabaceae, Brongniartieae), with description of two new local endemic species from Cuba

﻿Abstract Two new species of Harpalyce are described from Cuba, H.revolutasp. nov. from a serpentine area in the northern part of E. Cuba and H.marianensissp. nov. from calcareous areas in the southern part of E. Cuba. Both have relatively small flowers, with an up to 6 mm long standard, and 2–3 mm long wings. Harpalycemarianensis is further characterized by strongly suberous (corky) young branches of a spongy consistence, deeply furrowed longitudinally, and by leaflets covered with a particular type of orange, apparently disk-shaped, sessile glands abaxially. Harpalycerevoluta, moreover, has suborbicular or broadly elliptic leaflets with a strongly recurved or sometimes revolute margin, secondary veins inconspicuous on either side; the foliar glands, by their morphology and anatomy, are of a different type. An epitype is designated for the name Harpalyce and its type, H.formosa; the distribution of both new species and their close relatives is mapped, and an updated identification key is offered, to cater for all 16 currently recognised Cuban species.


Introduction
Harpalyce DC. is one of those genera that, so to say, sprouts new species on a regular basis, with apparently increasing frequency. The phenomenon is not limited to Cuba, but can also be observed in the two other, disjunct areas in which the genus occurs: Central Mexico to Nicaragua in Central America, and large parts of Brazil (São-Mateus 2018; São-Mateus et al. 2019;Rankin and González 2021). In Cuba, it is due to a variety of factors, among which are the occurrence in small populations of only a few individuals, long overlooked by collectors; and the fact that many collected specimens are incomplete, often lacking flowers and fruits, and unsuited for identification below the species group level, so that undescribed species often remain unrecognised in the herbaria.
In Cuba, from where 16 species (including the two present, new ones) are now known, Harpalyce was first collected by Charles Wright in the 1860s. The species described by Grisebach (1866) based on Wright's material remained the single Cuban one for well over half a century. In the 1920s, five more species were added, consequent to the intense collecting activities of the Swede E. L. Ekman and North American botanists based in New York. Cuban botanists recognised three more species, all described in 1950. Borhidi andMuñiz (1977) added no less than 5 species (not counting synonyms) based on the collections of local botanists kept in Cuban herbaria. Just recently, describing a new species of our own, we published an update of the genus for Cuba (Rankin and González 2021). Even then, we realised that our digest was by no means final, as several sterile specimens exist in Cuban herbaria, including the type material for species described by Borhidi and Muñiz (1977; see Rankin and González 2021) that cannot be placed reliably in the framework of the species so far described. Further field investigations are planned which, if successful, will yield complete, fertile material from the very same or nearby populations. Therefore, be warned: as soon as the so far incompletely known plants have been located and collected, further descriptions of new taxa will likely be forthcoming.

Materials and methods
The original and complementary literature dealing with Harpalyce (Fabaceae, Brongniartieae) was consulted. In addition to the ca. 170 materials studied earlier (Rankin and González 2021), about 80 specimens (including duplicates) of Cuban Harpalyce belonging to the herbaria B, HAJB, JE, and PAL-Gr were examined, not counting the digital images of other specimens (including types) available online and housed in relevant herbaria such as A, GH, GOET, HAC, HAJB, K, MO, NY, P, S, UC, and US (see also Rankin and González 2021 and its Appendix 1 [https://doi.org/10.3372/ wi.51.51204]). Herbarium codes follow Thiers (2022+). Georeferenced specimen label data were incorporated into the Flora Database of the Republic of Cuba (Greuter 2003); geographical coordinates were used to generate a distribution map, using the mapping software QIGS for Windows, version 3.16.
Leaf glands were studied using the stereo-zoom microscope system Olympus SZX16; the photographs were taken using the microscope's camera Olympus DP72 and its CellSens Standard software. Cross cuts of the lamina were obtained by razorcutting manually mature leaflets, previously soaked soap water at room temperature. Unstained cuts in aqueous solution were studied at 200× and 400× magnification by optical microscopy using a Carl Zeiss Axioscop instrument.

Notes on the type of and author citation for the name Harpalyce
Harpalyce formosa DC., Prodr. 2: 523. 1825. Holotype [or, if lectotype, designated as such by Arroyo (1976: 40)]: Original painting (#227) of "Astragalus Formosus. Sp. N." in the Candolle collection "Icones florae mexicanae" of the Sessé and Mociño paintings, and copies thereof, in the library of the Conservatoire botanique in Geneva (G). Epitype (designated here): Tehuacán, Puebla, June 1905, Purpus 1196). This specimen was revised by Arroyo (1976) and considered by her as the holotype (more likely it is the lectotype) of Harpalyce ferruginea Brandegee, a taxonomic synonym of H. formosa; it can be viewed on the Internet (https://global. plants.jstor).

Description of two new Cuban species of Harpalyce
Phenology. Collected in flower and with fruits in April and May. Etymology. Named after the type locality, Abra de Mariana. Distribution. Southern part of E Cuba, province Guantánamo, municipality San Antonio del Sur: Abra de Mariana and Baitiquirí. Grows in dry scrub vegetation on limestone substrate. Fig. 4.
Phenology. When collected in late April, the type material was in flower and with immature fruits. Some other specimens seen, collected in April as well, similarly were in flower and with immature fruits.

Etymology.
The epithet refers to the characteristic, revolute leaflet margin. Affinities and diagnostic features. Harpalyce revoluta, shows morphological affinities with H. marianensis (see above) and other Cuban species that have a relatively small (≤ 8-9 mm in length or diameter), orbicular to broadly elliptic standard and free keel petals, particularly with H. ekmanii, with which it shares glabrous or subglabrous leaflets; however, in H. ekmanii the leaflets are much larger (2.5-7.5 × 1.5-3 cm vs 0.6-2 × 0.5-1 cm), with flat (vs recurved to revolute) margins, and the sessile glands on the abaxial leaf face are fairly dense, less spaced. The size and shape of the standard relates H. revoluta to H. villosa, but the latter has leaflets that, particularly when young, are densely hairy abaxially, not glabrous except on the loosely hairy midvein, as in H. revoluta.

An update of the identification key for Cuban Harpalyce
In their recent survey of Cuban Harpalyce, Rankin and González (2021) presented a key for identifying the species they recognise. To account for the present, additional species, that key needs some modification. That key is reproduced here, with its 8 th dichotomy changed and expanded to account for the new species. Leaflets 5

Discussion
As we stated in the introduction, the two new species here described are likely not the last addition to the genus. There are incomplete specimens in the Cuban herbaria, HAC and HAJB in particular, that in all likelihood represent new, undescribed species, which we shall duly name as soon as we succeed in relocating them in the field and collect complete, fertile material; furthermore, in view of the species definition we here adopt, following the tradition established by other authors, any newly discovered Harpalyce population is likely to represent a new taxon. The question may be asked legitimately: is such a narrow species definition practical and defendable? We ignore whether our species are valid under a biological species concept. For all we know, they might all, or at least some of them, be freely interbreeding when brought together. Without much experimental work involving artificial crossing, it is impossible to know whether and to which extent evolutionary divergence has succeeded in establishing genetic barriers and cross-incompatibility between morphologically distinct populations. We are similarly ignorant of the extent of genetic isolation, if any, for many and indeed most species currently recognised in polymorphic Cuban genera. It is therefore legitimate to recognise morphologically distinguishable populations as separate taxa, and currently, no workable alternative to this course is available.
The role of small, isolated populations in evolution and speciation in higher plants has been studied by Runemark in the Aegean archipelago of Greece. He has coined the expression "reproductive drift" (Runemark 1969(Runemark , 1970 to describe and explain this phenomenon. His ideas have not received the attention that they deserve, and we would like to encourage others to apply them to understand and explain the polymorphism observed in Cuban species complexes with a comparable fragmented population structure which, in the case of Cuba, is usually associated with edaphic specialisation to small and scattered habitats which, for biological purposes, are comparable to archipelagos of small islands in a sea of generalistic environmental conditions. Concerning the general evolutionary context in which the Cuban populations are to be seen, it is useful to be mindful of the results and conclusions of recent phylogenetic studies (São Mateus 2018). That work was based on sequence analyses of both nuclear ribosomal and plastid DNA of an impressive sampling of taxa from Brasil (Harpalyce sect. Brasilienses Arroyo) and Mexico (H. sect. Harpalyce) but only a single one from Cuba (two accessions of, allegedly, H. cubensis Griseb., which in fact, however, both belong to H. suberosa Urb.) representing H. sect. cubenses Rydb.). By São Mateus' (2018) results, the genus Harpalyce and, presumably, its three geographically vicariant sections are natural, monophyletic taxa, with H. sect. Brasilienses being the sister clade of the two remaining sections, which are in turn sisters to each other (a rather shaky assumption, as long as a single Cuban taxon has been studied). The genus is old, assumed to be of Oligocene origin (> 30 Ma b.p.). Divergence of the northern sections from the Brazilian clade is also old (> 20 Ma b.p.). Species diversification, however, is fairly young, basically post-Miocene, an assumption that obviously, for Cuba, remains to be verified by the study of a larger number of taxa.
A further promising field of additional studies that we propose to follow up in the near future has opened up through the preliminary study of leaf glands of the new species here described. We found that the slight but obvious difference of the gland pattern they presented was based on obvious structural differences, to be observed on leaf transects, which have led to the definition of two distinct gland types that we have named "type A" and "type B". Screening the remaining species of the genus for the presence in them of these gland types, and of possible further types yet to be defined, is a promising field of study. Preliminary results of leaf surface studies indicate that the "type A" glands of Harpalyce marianensis are also found in H. ekmanii