|Species Days pictures|
Excerpts from and comments on a paper published by:
Loretta Goetsch, Andrew Eckert and Benjamin Hall
University of Washington
Classification of Rhododendron species based on morphology has led to a consensus taxonomy recognizing the major subgenera Azaleastrum, Hymenanthes, Pentanthera, Rhododendron, Tsutsusi and three minor ones.
To determine whether these subgenera are monophyletic and to infer phylogenetic relationships between Rhododendron sections and species, Goetsch et al. carried out a cladistic analysis using molecular data, including all groups within the genus.
For this purpose, they sequenced a large part of the nuclear gene RPB2-I, encoding a major RNA Polymerase II subunit, from 87 species and analyzed the data by maximum parsimony, maximum likelihood, and Bayesian methods.
The resulting phylogenies show subgenera Azaleastrum and Pentanthera to be polyphyletic and group all Rhododendron species (except the two in section Therorhodion) into three large clades.
Based upon these results, a modified system of Rhododendron classification is proposed, consolidating minor subgenera that are related and recognizing monophyletic subgenera.
More than 90% of the 1,025 Rhododendron species described prior to 1996 (Chamberlain et al. 1996) belong to the predominately Asian subgenera Hymenanthes, Rhododendron and Tsutsusi.
The first two of these have many species in the Himalayan- Southwest China region; in addition, the 300 species of section Vireya in subgenus Rhododendron are distributed mainly through the islands of the Malay Archipelago (Sleumer 1966), extending from their probable origin on the Asian mainland as far as northern Australia.
The geologically recent juxtaposition (< 10 million years ago) of the eastern and western halves of this archipelago raises interesting biogeographic questions for future phylogenetic study of Vireya species, as does the Himalayan orogeny (Irving and Hebda 1993) for Hymenanthes and Rhododendron species of the Sino-Himalayan area.
Rhododendrons of subgenus Tsutsusi have a mainly east Asian maritime distribution (Japan, Korea, Taiwan, and east China) with no species in either western Eurasia or North America.
Systematic studies that encompassed all sections and subgenera of Rhododendron were initiated by Sleumer (1949) who proposed a comprehensive system of Rhododendron classification in the form of a key to subgenera and sections.
Subsequently, the conclusions of a number of more narrowly focused morphological taxonomic studies (Sleumer 1968; Cullen 1980; Chamberlain 1982; Philipson and Philipson 1986; Judd and Kron 1995) were incorporated into an alternative Rhododendron classification.
This taxonomic system is now generally accepted by Rhododendron specialists (Cox and Cox 1997) because it embodies the findings of substantially all morphology-based Rhododendron systematic studies since 1980.
Significant differences between the Sleumer (1949, 1980) and Chamberlain et al. (1996) taxonomic systems concern subgenus Therorhodion, which Sleumer placed outside genus Rhododendron, and placement of the four species of section Sciadorhodion. Based on studies by Judd and Kron (1995), Chamberlain et al. (1996) assigned these species to subgenus Pentanthera, while Sleumer merged them with section Brachycalyx in subgenus Anthodendron, equivalent to subgenus Tsutsusi (Chamberlain and Rae 1990).
An interesting, although infrequently noted, feature of Sleumerís taxonomic key is the proximity of the deciduous section Pentanthera to the evergreen subgenus Hymenanthes. These taxa both lack lepidote scales and, for both, the new leafy shoots emerge from the axils of shoots from the previous yearís growth.
In subgenus Pentanthera the Chamberlain et al. (1996) classification system includes the major section Pentanthera, comprising 15 species from the southeast United States plus three from other regions: section Sciadorhodion and the smaller sections Rhodora (2 spp., North America) and Viscidula (1 sp., Japan). Other than having deciduous leaves covered in hairs and terminal rather than axillary inflorescences, few morphological attributes link these four sections together (Cox and Cox 1997).
Historically, the most taxonomically problematic rhododendrons have been the subgenera Azaleastrum, Mumeazalea and Candidastrum. Both classification systems place sections Azaleastrum and Choniastrum, which share the lateral inflorescence character, in subgenus Azaleastrum even though they differ consistently in number of stamens (5 vs. 10) and other characters (Philipson and Philipson 1986).
Because of distinctive floral and seed characteristics, the deciduous taxa R. semibarbatum Maxim. (Japan) and R. albiflorum Hook.f. (North America), were placed, respectively, in separate monotypic subgenera Mumeazalea and Candidastrum.
Two studies of molecular systematics across the genus Rhododendron have previously been published. The first used sequences from the chloroplast matK and trnk genes (Kurashige et al. 2001) and the second used nuclear ITS Sequences (Gao et al. 2002). As detailed in the Goetsch et al. paper, several of the contradictions between morphology-based Rhododendron taxonomy and the RPB2-I phylogeny determined in their paper are also evident in the plastid and ITS phylogenies, although those publications did not emphasize the contradictions.
In their investigation, Goetsch et al. recovered, sequenced and computationally analyzed sequences of RPB2-I from 87 Rhododendron species in order to address several related issues.
First, they set out to test whether the morphology-based sections and subgenera of Rhododendron proposed by the taxonomic systems of Sleumer (1949, 1980) and Chamberlain et al. (1996) are monophyletic.
A second objective was to resolve, irrespective of these and other taxonomic proposals, the relationships between all Rhododendron sections, including subsection Ledum and genus Menziesia (Kron and Judd, 1990).
The monophyletic groups so identified, together with morphological information, provide the basis for a revised classification system for Rhododendron, which is described briefly below.
The results of the Goetsch et al. investigation clarify the phylogeny of Rhododendron and suggest that several changes in the infrageneric systematics of Rhododendron are warranted.
Based upon the molecular data that they and others have obtained, certain revisions in the Rhododendron taxonomic system are proposed.
For taxa outside of subgenus Rhododendron, this system eliminates three subgenera and two sections that are present in the taxonomic system of Chamberlain et al.
Inclusion of section Pentanthera within subgenus Hymenanthes reflects the 100% bootstrap and Bayesian support for a clade containing only these taxa. Hymenanthes is monophyletic with 79% bootstrap support.
Sections Sciadorhodion and Viscidula and R. vaseyi (section Rhodora) from the discontinued subgenus Pentanthera are combined with sections Azaleastrum, Tsutsusi and Brachycalyx to form an expanded and revised subgenus Azaleastrum.
Sister groups in this subgenus are the sections Tsutsusi (largely evergreen) and Sciadorhodion (entirely deciduous).
While the RPB2-I phylogeny places section Choniastrum in clade A, as sister taxon to subgenus Rhododendron, Choniastrum lacks the attribute most characteristic of this subgenus, lepidote scales on the leaves.
For this reason, Goetsch et al. propose that Choniastrum be considered a separate subgenus.
Number of species analyzed, grouped by the taxonomic system of Chamberlain et al. (1996)
subgenus Azaleastrum Planch.
section Azaleastrum (2)
section Choniastrum Franch. (3)
subgenus Candidastrum Franch. (1)
subgenus Hymenanthes (Blume) K. Koch
section Ponticum G. Don (20)
subgenus Mumeazalea (Sleumer) W. R. Philipson & M. N. Philipson (1)
subgenus Pentanthera (G. Don) Pojarkova
section Pentanthera (5)
section Rhodora (L.) G. Don (2)
section Sciadorhodion Rehder & Wilson (2)
section Viscidula Matsum. & Nakai (1)
section Pogonanthum Aitch. & Hemsl. (3)
section Rhododendron (30)
section Vireya (Blume) Copel.f. (9)
subgenus Therorhodion (Maxim.) A. Gray (1)
subgenus Tsutsusi (Sweet) Pojarkova
section Brachycalyx Sweet (2)
section Tsutsusi (3)
Menziesia Smith (2)
Empetrum L. (1)
Dr. Hall presented and explained the data analysis behind this proposal at the 2005 Annual ARS Convention in Victoria, B. C. where he was a speaker in the program on April 29.