2  Materials and Methods

This study is primarily based on the Cuban and Antillean species of the genus Pleurothallis (sensu Luer 1986b). For palynological and phytogeographical studies, however, taxa from other pleurothallid genera were added to provide a broader view on these orchids. Pleurothallis ekmanii Schltr., Pleurothallis excentrica (Luer) Luer (bas. Octomeria excentrica Luer) and Pleurothallis ‘flabelliformis’ nom. prov. (bas. Octomeria prostrata H. Stenzel), three species of uncertain phylogenetic position, were treated under the generic name Pleurothallis, which, unfortunately, made the employment of an unpublished name unavoidable.

Material was collected in Cuba, Jamaica, the Dominican Republic, and Puerto Rico during the years 1997 - 2001. The vast majority of the samples was gathered during a 6-month stay in Cuba in 1998 (Tab. 1), another in 1999, and a third one of 3 months in 2000 in Jamaica and Puerto Rico. Initially, localities which should be visited were chosen from herbarium data and from information found in literature. Additional suggestions came from Dr. M. A. Díaz (HAJB), Dr. H. Dietrich (JE) and Dr. A. Urquiola Cruz (HPPR) resulting from former field trips within the Flora de Cuba Project (hereafter FCP).

	Fig. 2: Collecting areas in Cuba.
	1 – Guanahacabibes; 2 –Sierra de los Órganos; 3 – Cajálbana; 4 – Pan de Guajaibón; 5 – Taco Taco, Rangel; 6 – Montañas de Trinidad; 7 – Sierra de Banao; 8 – Turquino Massif; 9 – Guisa, Victorino; 10 – Gran Piedra; 11 – Palenque, Bernardo, Pico Galán; 12 – Sierra de Imías; 13 – Meseta del Toldo; 14 – Sierra del Cristal; 15 – Sierra de Nipe.

The majority of the more than 1300 vouchers collected belongs to the family Orchidaceae, with almost 500 representing taxa of the subtribe Pleurothallidinae. Collecting and transfer of the material took place in accordance with Cuban law and CITES criteria. In general, collecting of entire populations was avoided. In concordance with the rules of the FCP, herbarium material belonging to Pleurothallis will be split between HAJB and JE. [page 14↓]Holotypes and vouchers consisting of single specimens will be deposited in HAJB in any case. Material in spirit, CARNOY as well as silica dried samples belong to the personal collections of the author. All material that does not belong to Pleurothallis will be housed in HAJB.

Tab. 1. Survey of the field work in Cuba.

Besides these personal collections, material of the following herbaria was examined: AJBC, BSC, HAC, HAJB, HPPR, JE, B, BM, BR, DUKE, G, GH (AMES), GOET, IJ, K (including K-L), NY, S, UPRRP, and W. Xerocopies or scanned images of material of the following herbaria was included, too: B-W (IDC-copies), GH (AMES), M, S.

Collection of plant material

Plants were dried in a portable press in the field. To preserve the details of the delicate and fragile generative organs, part of the material was conserved in spirit.

As far as possible, each voucher was completed by information on

1. plant characteristics in vivo (colours, scents, growing habit, inflorescence and flower posture, flower characteristics);
2. population characteristics (number of individuals, ~ flowering individuals);
3. the locality (geographic position with GPS data), altitude, elevation above ground in the case of epiphytic populations (higher levels of trees etc. were examined by [page 15↓]eye and binoculars), cardinal point of the growing site on bark or rock, geological and edaphic characteristics;
4. bio-ecological data (phorophytes, accompanying flora, vegetation type, observations of potential pollinators).

Mature flowers for palynological studies were separately dried in paper envelopes.

Slide images were taken of the natural environment and the collected plants.

Samples for molecular analyses were taken from fruits as well as from young and mature leaves. They were externally cleaned with ethanol (70%), cut into pieces to facilitate dehydration and stored in perforated paper envelopes in silica gel. In the first run (1998) desiccated material was stored only at (subtropical) room temperature. Later, material was maintained under cool conditions as soon as the pieces were dry. Even small intervals of time (from one month on!) under subtropical conditions led to severe deterioration concerning quality and amount of extractable DNA.

2.2  Methods

Descriptions and drawings

Descriptions were made including all material examined. Since the majority of the species could be collected in flowering condition, flower sections were carried out based mostly on personal spirit collections of the author. In most cases type material was not dissected, for many of the unique historical collections by Jacquin, Wright or Schlechter, are poor in generative material. Dry herbarium material was rehydrated in boiling water. Spirit material was dissected without further preparation. Pencil drawings were later copied with ink and scanned at high resolution. Scanned images were then arranged in position and size and provided with scales. All drawings are done by the author.

Reproductive biology

Phenological information was gathered from observations in the field and from herbarium material.

Ecological and phytogeographical patterns

General phytogeography – Data of distribution was collected from herbarium sheets in the herbaria mentioned above and was completed by information from the following sources: Ackerman 1995, 1997; Adams 1971, 1972, & pers. comm.; Ames & Correll 1952; Carnevali & al. 2001; Correll 1965; Dix & Dix 2000; Dietrich s. p.8; Dod 1974, 1984b, 1986b, 1989b; Foldats 1970; Dressler 1993b; Garay & Sweet 1974; Gloudon & Tobisch [page 16↓]1995; Hamer 2001; Luer 1975a, 1975b, 1990, 1998b, 1998c, 1999a, 1999b, 2000; MacLeish & al. 1995; Nir 2000; Stehlé 1939; Williams 1951; Williams & Allen 1946; Williams & al. 1980. Additional information from the NYBG West Indian Orchidaceae Specimens database (http://www.nybg.org/bsci/hcol/wior/orchidchecklist.html) and the MBG W3TROPICOS database (http://mobot.mobot.org/W3T/Search/vast.html) was taken into consideration.
This analysis is aimed at the comparison of the Greater Antilles islands with adjacent Lesser Antilles and continental areas. It includes all species of Pleurothallis found in the Greater Antilles. Variables are: West Cuba, Central Cuba, East Cuba, Jamaica, Hispaniola, Puerto Rico, Lesser Antilles, Central America, and South America.

Vertical distribution – Data of the vertical distribution of Cuban Pleurothallis species was gathered from personal observations and herbarium material. Occurrence of taxa was recorded in altitudinal belts of 100 m. Gaps within the vertical distribution were interpolated, which did not alter the graph qualitatively.

Geology – The distribution of the Cuban species of Pleurothallis in vegetation on three distinct types of rock (limestone, magmatic or volcanic rock, serpentine) was analysed. Petrologic (type of rock) data was obtained from personal observations and information on herbarium sheets. These facts were compared with Borhidi (1996), Samek (1973). Serpentine, though representing only one of the ultrabasic rock types is used in the broader sense of the latter hereafter. In the case of volcanic rock, the data may comprise actually a variety of types, which could not be differentiated. In fact, this category is more a negative circumscription: in no case it does contain stands, neither on ultrabasic rock nor on limestone. These stands are limited to the Sierra Maestra chain.

Veget ation types – Data concerning vegetation types are based on personal observations and information on herbarium sheets. These were compared with Samek (1973), Borhidi (1996), and Capote & Berazaín (1984). Assigning epiphytes to the vegetation types suggested by these authors turned out to be problematic occasionally. First, herbarium data is very scarce. Pleurothallis is distributed almost exclusively in colline to montane areas, where vertically and horizontally interlocked vegetation types can change within a few meters. Information from herbarium sheets, apart from their second hand character, are often too general. Second, classification of vegetation types may differ substantially between authors. Three problems occurred:1) types were partially incomparable because of inconsistent methods used by the authors and due to inaccurate descriptions, 2) there were cases of general misapplication of terms, and 3) some types turned out to be useless in classifying epiphytes (Pleurothallidinae). The latter was detected only in the course of the present study and will be discussed in detail later.
In the present study vegetation types sensu (Borhidi 1996) were used, however, with [page 17↓]some changes. Considering the differences that exist in the usage of even such common terms as “elfin” and “montane” (Capote & Berazaín 1984; Borhidi 1996) it seems reasonable to define the terms used in this study.

• Elfin vegetation – dwarf forests, thickets and shrubwoods with cloud forest appearance (rich in different kinds of epiphytes; branches, stems and rocks covered by mosses and lichens). Found on isolated peaks and crests: Pan de Guajaibón; Sierra del Escambray: Pico Potrerillo, Lomas de Banao; Pico Cristal; Pico Galán; Turquino Massif: N-slopes, crests, and summits; summit of Loma del Gato; Gran Piedra. Includes rocky habitats (epilithic life forms).
• Montane rainforests – sensu Borhidi (1996) p.p. (includes some stands of Borhidi’s montane karstic forests). Sierra del Escambray (Pico Potrerillo, San Juan?), Sierra Maestra, Sierra del Purial.
• Montane rainforests on serpentine - sensu Borhidi (1996) p.p. (includes Borhidi’s submontane rainforests on serpentine in the Moa-Toa-Duaba area). Sierra de Nipe (lower W slope, EIIM), Sierra del Cristal (N-E-slope, lower W-slope), Moa, Toa, Duaba.
• Submontane seasonal rainforests – sensu Borhidi (1996) p.p. This includes the humid karstic forests of Oriente, which may be debatable. Therefore, the mogote complex is added further down as a separate unit. Vegetation of the mogotes in Oriente (Sierra de Nipe belt, N of Sierra Maestra, Yunque de Baracoa); vegetation in mesoclimatic niches in Pinar del Río and (hoyos and below shaded slopes), Sierra Maestra between 200-800 m.
• Semi-deciduous mesophytic forests – sensu Borhidi (1996). Mogotes (surrounding forests).
• Dry (microphilous) evergreen forests – sensu Borhidi (1996). Coast of Guanahacabibes peninsula.
• Dry (thorny) lowland serpentine shrubwoods (cuabal) – sensu Borhidi (1996). Cajálbana, Canasí.
• Semi-dry (semi-thorny) serpentine shrubwoods (charrascal) – sensu Borhidi (1996) p.p. (comprises Borhidi´s semi-dry lowland and montane shrubwoods). East Cuba: lowlands and hills N of Sagua-Baracoa; summit and SE+N-slope of La Mensura; S-slope of Sierra del Cristal 700-1000 m; Toldo massif; Alto de la Iberia.
• Karstic forests – although this is a mosaic of interlocked forest types already included in this list, I think it is important to include this traditional Cuban vegetation complex as a separate unit. It comprises the mogote systems of West, Central and East Cuba.[page 18↓]
• Xerothermic Pinus cubensis forests on serpentine – sensu Borhidi (1996). East Cuba: serpentine areas of northern and eastern Oriente.
• Pine forests on limestone – sensu Borhidi (1996). Occurs only on the edge of pine forests on serpentine in Monte Cristi Borhidi (1996)
• Montane pine forests – sensu Borhidi (1996): isolated patches in the Turquino region and Gran Piedra.
• Gallery forests and shrubwoods – the term (hereafter as gallery forests) comprises colline, submontane and montane river banks and creeks. It should be stressed that headwaters are included.
• Secondary forests – the term appears only in the Spanish summary of Borhidi (1996). In the present study it refers to secondary mogote vegetation (stands of former seasonal evergreen forests) and cupeyales in the Sierra de Nipe (originally montane rainforests on serpentine).
• Secondary thickets – refers to a Gleichenia(?) – Dicranopteris – treefern shrubwood (Borhidi 1996: matorrales seminaturales) on the summits of Los Calderos (Sierra de Imías).
• Cultivations – coffee plantations are the only cultivations where species of Pleurothallis have been found so far.

Computing and representation of the data in diagrams

Dendrograms – Data was set up in matrices in which species were coded as present (1) or absent (0). Species represented by single collections (Pleurothallis appendiculata, P. ghiesbreghtiana, P. aristata) or questionable classification (P. murex was described from sterile material) are not included in the matrices that were used to show correlations. To assess and show groupings of variables, Maximum Parsimony (MP) analyses were conducted. Dendrograms were initially rooted with a hypothetical outgroup with all species coded absent (Lundberg rooting), which was abandoned in a second run (see p. 78). Computation was carried out with the PHYLIP package (Felsenstein, J. 1989: PHYLIP (Phylogeny Inference Package) version 3.5c). This algorithm follows largely PAE (parsimony analysis of endemicity) introduced by Rosen & Smith (1988). See Trejo-Torres & Ackerman (2001) for a detailed discussion of this method.

3-dimen sional diagrams – Most ecological variables, e.g. types of rock, vegetation, and vertical amplitude, were examined as to their correlation with overall horizontal distribution of the taxa. Although the same data was statistically evaluated, too, 3-dimensional diagrams were chosen for presentation, because of the intriguing and lucid way they convey information. In order to present the diagram in a clearly arranged manner, species [page 19↓]were classified according to their occurrence in n locations (i.e. frequency) in the Caribbean (6 locations: West Cuba, Central Cuba, East Cuba, Jamaica, Hispaniola, Puerto Rico) and adjacent areas (GA-SA: Lesser Antilles, Central America, South America). The latter, because of its vast area, does not represent a 7^th location. Thus, species of distribution outside the Greater Antilles, may or may not be present in all 6 inner Antillean localities. Again, in order to avoid overloading the diagrams, ecological amplitudes were classified, too, usually in three units representing the growing amplitude. Hence, horizontal distribution is growing on the x-axis from left to right and the ecological amplitude from the foreground towards the background (see Fig. 56 e.g.).

Statistics (correlation) – As just mentioned, an association between horizontal distribution and the ecological amplitude was examined. Species were grouped according to the number of areas they occur in (see previous paragraph for a definition). All correlations were performed with the SPSS 10.0 package, r_s represents the Spearman’s rank correlation coefficient (Sokahl & Rohlf 1995), N the number of species, P is the significance value. The significance criterion for all tests was set at alpha=0.05, therefore P-values smaller than P=0.05 are considered as significant.

Palynology

The material included in the study of pollen morphology comprises ~100 samples of 72 pleurothallid and 2 other orchidaceous species (Tab. 2). The latter were used in the molecular study by (Pridgeon & al. 2001) as outgroup taxa. All but three Cuban species (Pleurothallis caymanensis, P. oricola, P. nummularia) currently treated as Pleurothallis could be included. With the exception of Brachionidium parvum and Pleurothallis aristata, from Puerto Rico, all Cuban taxa were represented by one local voucher at least. The Cuban endemic Pleurothallis grisebachiana, a widespread species variable in size and flower coloration, is represented by 10 samples from 9 different localities (Fig. 45), to test the degree of variation in pollen morphology. The same applies to Pleurothallis mucronata (Fig. 46) and P. ekmanii (Fig. 47) three Cuban localities each), P. ghiesbreghtiana (three samples from Central America, and one each from Cuba, Jamaica and Puerto Rico) and Pleurothallis ruscifolia (Jacq.) R. Br. (5 samples). The only subgenus of Pleurothallis that is endemic in the Antilles (~10 species, Luer 2000), subg. Antilla Luer, is represented by 7 species with 3 being restricted to Cuba.
Antillean material from closely related genera (Brachionidium, Lepanthes, Lepanthopsis, Platystele, Stelis, Trichosalpinx, Zootrophion) as well as from presumed sister or outgroup taxa in previous cladistic studies of Pleurothallidinae, Arpophyllum (Neyland & al. 1995, Pridgeon & al. 2001) and Dilomilis (Pridgeon & al. 2001) was included for comparative purposes.

All dry specimens were rehydrated for 1-5 min in hot water. All spirit material (Stenzel 482, Stenzel 765, Stenzel 1298) was dissected directly. Pollinia were removed, mounted on double tape and dried at room temperature. Pollinia were finally gold sputtered 100 s at 30 mA and a distance of 6 cm from the electrode.

Pollinia were examined with a JEOL SM 6300 SEM at the Palynological Laboratory of the Natural History Museum Stockholm and at the Museum of Natural History Berlin. Working data is as follows: Acceleration voltage 15 kV, probe current 3x10^-11 A, distance 8-12 mm. Since the upper half of the pollinium tends to be more stable (Schill & Pfeiffer 1977) and generally free of elastoviscin (Stenzel 2000), data of this area was used for comparison. Caudicular regions were documented as well. If not indicated otherwise, photos show the entire pollinarium and details on the side of the upper pollinium which faces the tapetum. Pollen terminology, if adequate, was applied according to Punt & al. (1994). For description of additional terms see glossary (p. 170).

[page 21 - 22↓]

Tab. 2: Plant material used for palynological studies (species present in Cuba in bold type).

* – CR = Costa Rica, CU = Cuba, DR = Dominican Republic, EC = Ecuador, HA = Haiti, JA = Jamaica, NI = Nicaragua, PR = Puerto Rico, SU = Surinam, VE = Venezuela. ** – Gr = Prov. Granma, Gu = Prov. Guantánamo, Ho = Prov. Holguin, PR = Prov. Pinar del Río, SC = Prov. Santiago de Cuba, SS = Prov. Santi Spíritus.

[page 23↓]Genetics

Plant material used in this study is listed in Tab. 3. All but 2 of the 39 Pleurothallis species (Pleurothallis appendiculata, P. murex) currently known to occur in Cuba could be included, i.e. the range is limited geographically. One of the 33 species successfully sampled could not be recollected in Cuba, i.e. material from other islands had to be employed (Pleurothallis aristata from Puerto Rico). Two species are represented by two accessions each. P. trichophora with material from East and Central Cuba which marks the phytogeographical limits of this plant, and P. ghiesbreghtiana with one voucher from West Cuba and the other from Puerto Rico.

Tab. 3: Plant materials used in molecular studies.

* – CU = Cuba, HA = Haiti, PR = Puerto Rico. ** – Gr = Prov. Granma, Gu = Prov. Guantánamo, Ho = Prov. Holguin, PR = Prov. Pinar del Río, SC = Prov. Santiago de Cuba, SS – Prov. Sancti Spíritus.

Material – Since it was not possible to extract DNA from existing herbarium collections, material for molecular investigations had to be collected in the wild. Outside ITS sequences for outgroup comparison were kindly provided by Dr. A. M. Pridgeon. They were originally published by Pridgeon & al. (2001) and will be cited hereafter, if necessary, under the GenBank accession number.

DNA extraction – DNA was extracted at Friedrich-Schiller-University in Jena from 0.01-0.5g of dried leaves or fruits, following a modified 2%-CTAB (hexadecyltri-methylammoniumbromide) procedure of Doyle & Doyle (1987) and Hellwig & al. (1999). DNA was purified using Qiagen tip-20 columns (Qiagen Inc., Hilden, Germany) following the manufacturer’s protocols. Purified DNA extracts are stored in HUB.

Amplification – All PCRs were run using Taq DNA Polymerase Kit from Qiagen (Qiagen Inc.). Initially, Baldwin’s (1992) primers P1 (=”ITS5” sensu Baldwin), P2 (=”ITS2”), P3 (=”ITS3”), P4 (=”ITS4”) were employed. They yielded two fragments. One was the result of primers P1 (5'-GGA AGT AAA AGT CGT AAC AAG G-3') and P2 (5'-CTC GAT GGA ACA CGG GAT TCT GC-'3). It covered ITS1 and 2/3 of 5.8S gene with a total length of ~380 bp. The other fragment of ~440 bp included 2/3 of the 5.8S gene and ITS2. It was the product of primers P3 (5'-GCA TCG ATG AAG AAC GCA GC-'3) and P4 (5'-TCCTCCGCTTATTGATATGC-'3). Both fragments are overlapping in the 5.8S region, so by combining the two, the complete ITS1-5.8S-ITS2 (hereafter ITS) sequence could be achieved.

Tab. 4: Pleurothallis gemina: 17SE (p.p.), ITS1, 5,8S, ITS2, and 26SE (p.p.) sequences.

Noncoding (ITS) regions in lower case, primer target regions underlined, primer pair P1/P2 in red, primer pair P3/P4 in blue, amplification direction of primers indicated by „<“ and „>“. Base mismatches in primers in bold type.

Tab. 4 illustrates the position of each primer within the nrDNA region of interest. Baldwin’s primers gave ad hoc excellent results in ITS1 (P1-P2) but less so in the case of ITS2 (P3-P4). Adjusting of the annealing temperature did not improve the quantity and quality of the products. Since the PCR with P1-P4 gave the same blurred PCR products as P3-P4, I [page 25↓]assumed that the P4 must include bases mismatching the 26SE target region. However, this cannot explain the notorious difficulty amplifying ITS2. Rather than base mismatches, secondary structures could explain those problems, for later usage of P4 in nested cycle sequencing with templates gained from non-P1-P4 PCR, showed no problems at all.
Sequence analyses revealed a three base mismatch in primer P2 compared with the actual 5.8S target region in all species examined, whereas primer P3 has one base mismatch (Tab. 4). Cycle sequencing functioned well despite these incongruities.

In those cases where PCR with Baldwin’s primers failed, a second set of oligos was employed (Sun & al. 1994). Using the primers “17SE” (5'-ACG AAT TCA TGG TCC GGT GAA GTG TTC G-3') and “26SE” (5'-TAG AAT TCC CCG GTT CGC TCG CCG TTA C-3') PCR yielded a fragment of ~ 900 bp comprising ITS1, 5.8S gene, and ITS2 along with ends of the bordering coding regions. Since these oligos anneal further outwards from primers P1 and P4, P1-P4 could be used for nested cycle sequencing.
The amplification was conducted via the following thermal cycle profile: 150s95°C/ 28-34x [30s95°C/60s50-58°C/ 60s72°C]/ 300s72°/ 4°C. PCR products were purified using the QiaQuick® PCR Purification Kit (Qiagen Inc.).

Cycle sequencing – Both strands of the amplified gene fragments were directly cycle sequenced in 10 µl volumes containing 2 µl of ABI Prism BigDye Terminator cycle sequencing reaction mix (ABI Biosys-tems), 0.5 µM primer, 2 µl dd H2O and 4 µl DNA. Sequencing products were purified following the ABI standard protocol adjusted to 10 µl reaction volume by addition of dd H₂O. Sequences were obtained by running the sequencing products on an Perkin Elmer ABI 377 automated sequencer. The resulting sequence electropherograms of both strands were corrected manually for misreads and merged into one sequence file using BioEdit (version 5.0.9.; Hall 1999) for Windows 95, which was generally used to store and manage sequences.

Alignment and phylogenetic analysis – The question of homology is a crucial problem in inferring phylogenetic relationships from nucleotide sequence data (Doyle & Davis 1998), i.e. the insertion of gaps is one of the most critical steps in molecular analysis. To compare the influence of the extent of indels, sequences were initially aligned using CLUSTAL W (Thompson & al. 1994) employing different delay values and gap costs and then adjusted by eye (see appendix for sequences). Unambiguous indels that seem to be duplications of upstream sequences (e.g. ATAT à ATATAT) were coded as single events. From each matrix an additional set with ambiguous indel regions being excluded was drawn. A fifth matrix was created by stripping an aligned set of all columns containing gaps. Cladistic analysis was conducted with these 5 data sets using sequence AF262915 (Dilomilis montana (Sw.) Summerh.) from (Pridgeon & al. 2001) as outgroup.
Substitution patterns were analysed by comparing observed and expected substitutions [page 26↓]among all sequences. The number of expected substitutions was calculated from base frequencies. Substitution saturation was evaluated by plotting portions of transitions and transversion against distance.
Phylogenetic trees were built using parsimony algorithms (Maximum Parsimony – MP hereafter). Computation was accomplished with PAUP* version 4.0b8 (Swofford 2000). The following settings were chosen: Fitch parsimony (Fitch 1971), heuristic search, gaps as fifth character, 100 replicates of random sequence addition, tree bisection-reconnection (TBR) branch swapping, MulTree.

Relative support of the topologies found was evaluated with 1000 bootstrap subsets of each of the 5 matrices.

Sequence editing and statistical analysis were performed using BioEdit and DAMBE (version 4.1.19; Xia & Xie 2001).

Taxonomy

Although profoundly challenged by molecular data (Pridgeon, Solano & Chase 2001, Pridgeon & Chase 2001) the morphological concept of Luer (1986c) will be the basis of this work. The main reasons are

1. less than 60 % of the supraspecific taxa proposed by Luer and included in that molecular study were represented by the type species.
2. Pridgeon & Chase (2001) based their ~500 nomenclatural transfers on the sequences of 187 pleurothallid specimens. Those epithets not sequenced were transferred merely on the basis of Luer’s morphological system. A great number of taxa of Luer’s system, however, had been shown to be paraphyletic in the same molecular study.
3. 15 of the 39 Cuban species were transferred to other genera, however, only 3 species were effectively included in the molecular screening.

There are more reasons why to handle the new system with extreme care. In order not to anticipate the results of this work they will be discussed later (p. 112).

General rules

Authority abbreviations follow those of Brummit & Powell (1992) and abbreviations of herbaria are in concordance with those of the Index Herbariorum (http://www.nybg.org/bsci/ih/searchih.html). Name of the Cuban provinces were abbreviated according to the rules of the FCP (Greuter & al. 2002): PR – Pinar del Río, SS – Santi Spíritus, Ci – Cienfuegos, Ho – Holguín, Gr – Gránma, SC – Santiago de Cuba, Gu – Guantánamo.