systematics of Valerianaceae

Evolutionary biologists and ecologists have long been interested in explosive radiations of species that exhibit a great degree of morphological, or phenotypic, diversity over a wide range of habitats. Such lineages are prime candidates for the studies of adaptive morphological radiations and, in the presence of reliable fossil evidence, the study rates of speciation. The asterid family Valerianaceae (Dipsacales) is comprised of about 300-350 species that are placed into 14-17 genera (Graeber, 1906; Meyer, 1951; Backlund, 1996). This group is characterized by its herbaceous growth habit, opposite leaves (rarely alternate), small, sympetalous and often asymetric flowers, and the presence of a variety of valepotriate compounds.

Members of this group occur in a wide variety of ecological settings throughout the Northern Hemisphere and the Andes of South America, where they are often found in mid to high elevations. The genera Triplostegia and Nardostachys are found exclusively in themid to high elevations of the Himalayas. Likewise, several species of Valeriana andPatrina are endemic to the high mountains of Asia. Throughout their range in South America, the species of Valerianaceae have a high occurrence in the paramo. In South America, the paramo is scattered along the crests of the highest ranges in the northern Andes or on isolated mountaintops from 3000 to 5000 meters and is divided into several vegetation zones, defined by altitude, the soil types, and the plant growth forms found. The species of Valerianaceae occur in all of these vegetational zones and exhibit a wide range in growth forms, from rosette plants (Valeriana rigida),microphyllous shrubs (V. microphylla), to an annual vine-like species (V.chaerophylloides).

Although little phylogenetic work has been done at the family level, directly, the relationship of Valerianaceae within the Dipsacales has been investigated quite extensively using morphological data (Judd et al., 1994; Backlund and Donoghue, 1996). Within the Dipsacales, these phylogenetic analyses place Valerianaceae as sister group to Dipsacaceae. Both of these herbaceous groups are united by having distinctive pollen morphology and chlorophyllous embryos, a trait unique within Dipsacales (Backlund and Donoghue, 1996). The link between Valerianaceae and Dipsacaceae is further supported by other morphological characters, including; simple vessel perforations, modification of calyx lobes, and the reduction in the amount of endosperm (Judd et al., 1994; Manchester and Donoghue, 1995; Backlund and Donoghue, 1996).

Various molecular data sets, including sequence data from the chloroplast genes rbcL (Donoghue et al., 1992) and ndhF (Pyck et al., 1999), restriction site data (Downieand Palmer, 1992), as well as combined morphological and molecular data (Backlund and Donoghue, 1996; M. J. Donoghue et al., ms; Bell and Donoghue, ms) have been collected to investigate the relationships within Dipsacales. These molecular phylogenetic studies also support the close relationship between Valerianaceae and Dipsacaceae, with this Valerianaceae-Dipsacaceae group being sister to the Morinaceae, another herbaceous dipsacalean group. Preliminary evidence from additional chloroplast markers (trnL-F IGS region, ndhF, matK- Bell and Donoghue,ms) do not ssupport the grouping of Triplostegia with Valerianaceae, but insteadplaces in a basal position of a clade containing the taxa of Dipsacaceae.

Currently, no phylogenetic analysis has included a complete sample of the genera of Valerianaceae, much less an extensive sampling of the larger and more diverse groups, such as Valeriana, Valerianella, and Patrinia. In the proposed study I am planning to:

• infer the phylogenetic relationships of Valerianaceae and closely related groups using several molecular markers
• examine patters of morphological evolution in Valerianaceae, particularly the recurrence of the reduction of floral parts and occurrence of asymmetric flowers
• compare rates of speciation in different lineages within Valerianaceae and compared to other lineages of Dipsacales as will as other angiosperm groups
• examine the age and biogeographic history of this group

Background Information on Valerianaceae

Originally described by Linneaus (1753, 1754), Valerianaceae is currently subdivided into three tribes (Graeber, 1906; see table 1): 1) Patrinieae, 2) Triplostegieae, and 3) Valerianeae. Many authors (Backlund, 1996) believe that Valerianaceae is comprised of 14 genera, two in the tribe Patrinieae (Patrinia andNardostachys, a single genus (Triplostegia) in the tribe Triplostegieae, and the 11remaining genera are grouped into five subtribes in the Valerianeae. Recent treatments of the South American taxa have argued for grouping most of these species in the genus Valeriana, reducing the number of genera within the family toeight(Larsen, 1986; Eriksen, 1998).

Flowers in Valerianaceae are small, often asymetric, and perfect to gynodioecious. The corolla is usually five parted, but there are many cases of reduction and fusion of floral parts. Flower morphology in Valerianaceae is quite disparate. Although most corollas are pentamerous, there are several groups that have a reduction to four (Phyllactis and Aretiastrum) and even three (Valeriana, Phyllactis, and Belonanthus). Most species within the family have a characteristic gibba (nectar container) which is fairly minute, however, several species (Centranthus and Plectritis) have evolved amore elaborate spur, presumably an adaptation to a specific pollinator. Although most treatments list Valerianaceae as entomophylous, little to nothing is know about pollination in this group.

Stamen number is often reduced, varying from one to five. The number is usually constant within genera and is often a reliable and useful taxonomic character. In Patrinia, stamen number can vary from one to four, or rarely five. Likewise, thepresence or absence of a calyx is highly variable across the family. When present it is usually uniform in appearance within a genus, the calyx segments are either leaf- or tooth-like (Eriksen, 1989) or pappus like. Ovaries in Valerianaceae are inferior and basically three-carpellate with only one of the carpels developing in the mature fruit. The sterile locules range from being extremely reduced to large and inflated. The presence or absence of sterile locules is presumably correlated with means of dispersal and ecology. In addition to the reduction/inflation of the sterile locules, various modifications of the calyx and seed morphology has been considered to be a useful taxonomic character at the generic and species level.

Although many authors would agree that species of Valerianaceae represent a monophyletic group, there has been some question as to the relationship of Triplostegia. The phylogenetic position of this small Asian genus has long beendebated; it has been placed in both Valerianaceae and Dipsacaceae, and more recently in the monogeneric Triplostegiaceae (Airy Shaw, 1964; Thorne, 1992; Peng et al., 1995). Recent analyses (Bell and Donoghue, ms) based on additional chloroplast sequences (matK, trnL-F IGS region, and ndhF) lend support for Triplostegia beingallied with Dipsacaceae. All three markers support this relationship and a combined analysis with these markers and rbcL sequences provide strong support (bootstrapvalue of 100 %) for this grouping.

The species of Valerianaceae, together with the Dipsaceae and Morinaceae, form the lineage Valerina (Donoghue, et al. ms) The supernumerary bracts present in the Linnaeeae (of the traditional Caprifoliaceae) have been hypothesized to be a morphological precursor to the epicalyx found in Morinaceae and Dipsacaceae (Hoffman and Gottmann, 1990; Roels and Smets, 1996). This distinctive morphological feature presumably evolved in a common ancestor of Morinaceae, Valerianaceae, and Dipsacaceae, with subsequent loss in Valerianaceae (with the exception of Triplostegiaand Patrinia).It has been interpreted that the wing-like structure on the fruit ofPatrinia is homologous with the supernumerary bracteoles in Linnaeeae or possibly amodification of the epicalyx in Morinaceae and Dipsacaceae. A detailed morphological/developmental study has yet to be undertaken to assess this homology.

Valerianaceae are thought to have originated in Asia (Hock, 1901; Erickson, 1989). Recent phylogenetic analyses based on both molecular sequence data from the chloroplast gene rbcL (Donoghue et al., in press), ndhF (Bell and Donoghue,unpublished data), morphology, and combined evidence (Donoghue et al., manuscript; Bell and Donoghue, unpublished data), place the genera Patrinia, andNardostachys (both native to eastern Asia) at a basal position within the family (withtheexclusion of Triplostegia). The remaining genera within Valerianaceae can bedivided into two separate biogeographical realms: 1) an Old World assemblage consisting of the genera Centranthus, Fedia, Valeriana, Valerianella, and 2) a New Worldgroup represented by 8 genera, including; species of Valeriana, Plectritis, Phyllactis,Aretiastrum, Belonanthus, Phuodendron, Astrephia, and Stangea. Within the New WorldValerianaceae, the majority of the species occur in the Andes of South American. Of the 300 species of Valerianaceae, nearly 200 of them are found to South America, with a high degree of endemism: 35 species (8 endemic) in Ecuador; 42 species (13 endemic) in Chile; 90 species (55 endemic) in Peru, 48 species (15 endemic) in Argentina;; and two species in Venezuela (zero Endemic). Recent treatments of the South American taxa have argued for grouping most of these species in the genus Valeriana, reducing the number of genera within the family to eight (Larsen, 1986;Eriksen, 1998).

Throughout its range in South America, the species of Valerianaceae have a high occurance in the paramo. The paramo is a vegetation type that occurs between the upper limit of the contiguous forest and the upper limit of plant life (i.e. snow pack) which is characterized by tussack grasses, large rosette plants, evergreen-sclerophyllous shrubs, and cushion plants. In South America, the paramo is scattered along the crests of the highest ranges in the northern Andes or on isolated mountaintops from 3000 to 5000 meters and is divided into several vegetation zones, defined by altitude, the soil types, and the plant growth forms found. The species of Valerianaceae occur in all of these vegetational zones and exhibit a wide range in growth forms, from rosette plants (Valeriana rigida), microphyllousshrubs (V.microphylla), to an annual vine species (V. chaerophylloides). Assuming that the SouthAmerican species of Valerianaceae have come from the north, which is reasonable since the family is absent from the southern hemisphere except South America. Eriksen (1989) hypothesized that the South American Valerianaceae represent three separate invasions. He based his hypothesis on morphological features (the number of anther locules) shared by distinct groups found in South America. One invasion presumably gave rise to the mainly Andean subgen. Phyllactis, with the closing of theIsthmus of Panama as a possible pathway; another colonization involving the very distinct sect. Porteria in Venezuela, presumably using the Caribbean Islands asstepping stones; and athird and final invasion by a single species, Valerianachaerophylloides (sect. Astrephia), also across the Isthmus.

Van der Hammen (1979), working on the palynological records from the Eastern Cordillera of Colombia, reports Valeriana pollen among the elements of the primitiveparamo flora originating in the upper Pliocene and lower Pleistocene. This evidence seems to support the hypothesis that Valeriana migrated to South America after theformation of the Isthmus of Panama which took placeabout 5.7 million years ago (Simpson, 1979). The upheavel of the Northern Andes took place mainly in the late Pliocene (Van der Hammen, 1979). The possibility of a large number of new habitats has been suggested to have led to an extreme diversification of Valeriana (Eriksen,1986) which now consist of herbs and shrubs, scandent shrubs, cushion and acaulescent rosette plants and even reports of small trees and rosette plants with long trunks (Eriksen, personal obs.). The presence of fossil data in Valerianaceae and related groups (Simpson, 1979; Manchester and Donoghue, 1996; Backlund, 1996) allow us to explicitly test hypotheses of biogeographic origins as well as rates of diversification. Fossils data can provide calibration points to be used along with molecular sequence data, and a phylogeny, to estimate divergence dates. These divergence dates can then be used to estimate absolute rates of diversification to investigate hypotheses of adaptive radiations or key innovation hypotheses. Although Valerianaceae does not have rich fossil record, fossils have been documented for the family and closely related groups. The distinctive fruits of Patrinia have been documented from the Miocene to Plioceneof Poland and Russia (Lancucka-Srodoniowa, 1997)as well as from the Miocene of Japan (Ozaki, 1980). Likewise, Valeriana is known on the basis of fossil fruits from thelate Miocene and Pliocene in Europe (Mai, 1985).

Some literature on Valerianaceae and Dipsacales

Airy Shaw, H. K. 1964. Diagnoses of new families, new names, etc., for the seventhedition of Willis's Dictionary'. Kew Bull. 18:249-271.

Backlund, A. 1996. In A. Backlund: Phylogeny of the Dipsacales, Doctoral thesis insystematic botanyfrom Uppsala University, Sweden.

Backlund, A. and Donoghue, M. J. 1996. Morphology and phylogeny of the Dipsacales. In A. Backlund: Phylogeny of the Dipsacales, Doctoral thesis insystematic botany from Uppsala University, Sweden.

Donoghue, M. J. Olmstead, R. G., Smith, J., F. and Palmer, J. D.1992. Phylogenetic relationships of Dipsacales based on rbcL sequences. Annals of the Missouri Botanic Gardens79:333-345.

Downie, S. R. and Palmer, J. D. 1992. Restriction site mapping of the chloroplast DNA inverted repeat: a molecular phylogeny of the Asteridae. . Annals of the Missouri Botanic Gardens 79:266-283.

Eriksen, 1989. Note on the generic and infrageneric delimitation in the Valerianaceae. Nordic Journal of Botany 9: 179-187.

Graeber, P. 1906. Valerianaceae andinae. Bot. Jahrb. Syst. 37: 436-451.

Hoffman, U. and Gottman, J. 1990. Morina L. and Triplostegia Wal ex DC. ImVergleich mitValerianaceae und Dipsacaceae. Bot. Jahrb 111: 499-553.

Larsen, B. B. 1986. A taxonomic revision of Phyllactis and Valeriana sect Bracteata (Valerianaceae). Nordic Journal of Botany 6:427-446.

Linnaeus, C. 1753. Species plantarum. Stockholm.

Linnaeus, C. 1754. Genera plantarum. Stockholm.

Manchester, S. R. and Donoghue, M. J. 1995. Winged fruits of Linnaeeae (Caprifoliaceae) in the Tertiary of Western North America: Diplodipelta gen. Nov.International Journal of Plant Science 156: 709-722.

Meyer, F. G. 1951. Valeriana in North America and the West Indies (Valerianaceae). Ann. Miss. Bot. Gard. 38: 377-503.

Peng, C.-I. Tobe, H. and Takahashi, M. 1995. Reproductive morphology and relationships of Triplostegia (Dipsacales). Bot. Jahrb. Syst. 116: 505-516.

Pyck, N. Roels, P. and Smets, E. 1999. Tribal relationships in Caprifoliaceae: evidence from a cladistic analysis using ndhF sequences. Syst. Geogr. Pl. 69: 145-159.

Roels, P. and Smets, E. 1996. A floral ontogenetic study in the Dipsacales. International Journal of Plant Sciences 157: 203-218.

Simpson, B. B. 1979. Quaternary Biogeography of the High Montane Regions of South America. In Duellen, W. E. (ed.), The South American Herpetofauna: its origin, evolution, and dispersal. Univ. Kansas Mus. Nat. Hist. Monogr. 7: 157-188.

Thorne, R. F. 1992. Classification and geography of the flowering plants. Botanical Review (Lancaster) 58: 225-348.

Van der Hammen, 1974. The Pleistocene changes of vegetation and climate in tropical South America, Journal of Biogeography 1: 3-26.