Finally a Family Name: Hostaceae!
W. George Schmid, Tucker, Georgia

Page 2

   Most classic texts, as for example Baker (1870), place Hosta in the family Liliaceae (lilies) usually in association with genus Hemerocallis (daylilies). For over a century placement with the family Liliaceae has been the most accepted taxonomic position on the familial level (Gray 1950; Hylander 1954; Hutchinson 1964). No serious objections were raised against the earlier placements until the 1930s, when cytologists found Hosta has a Karyotype similar to Agave and Yucca. Sato (1942) found that the Yucca-Agave complex karyotypically similar to that of Hosta and suggested evolutionary affinity between the two. These findings gave rise to placing the genus Hosta in the family Agavaceae (Moran 1949). The Karyotype of a given genus is determined by the study of cells and cell content, which includes the number and appearance of the chromosomes in a cell, in our case that of a Hosta species. These arrangements found support among cytologists, because karyotypical similarities appear to be common among plants formerly placed in the Liliaceae. However placements based on cytological evidence alone are no longer considered conclusive. Kubitzki (2010) published a complete revision of the taxonomic placement of Hosta. He assembled data from many scientific sources. His close examination is the most detailed among all of the published taxonomic and phylogenetic appraisals. He has published 9 volumes of over 500 pages each, and Hosta is included in Vol. III of The Families and Genera of Vascular Plants Flowering Plants; Monocotyledons: Lilianae (except Orchidaceae) by Klaus Kubitzki, H. Huber, P.J. Rudall, P.S. Stevens, and T. Stülzel of the Hamburg Institut für Allgemeine Botanik (HBG). Kubitzki detailed the Karyotype of the genus Hosta and his thorough assessments places Hosta in the family Hostaceae as proposed and validly published by Mathew (1988).

Methodus Plantarum Sexualis in sistemate naturae descripta (1736)
The page shows the Linné sexual classification system (24 classes).
Drawn by Georg Dionysius Ehret who wrote the names in the right edge.

   The data assembled by Kubitzki supports the latest systematic position of the genus Hosta in this monotypic family. Kubitzki’s data analysis shows that the chromosome numbers reported in Hosta (with 20-30 species examined) are based on x = 30 with 2n = 60 for the majority of the species (Sen 1975). The Karyotype (number, size, and shape of chromosomes) of Hosta is trimodal with 4 pairs of large, 2 pairs of medium and 24 pairs of small chromosomes. On the other hand, Agave has a bimodal Karyotype with 2n = 60 with 10 pairs of large and 50 pairs of small chromosomes (Tamura 1995). The large chromosomes of Hosta cells show a number of localized chiasmata, which are the points where two homologous chromatids exchange genetic material during meiosis. These chromosomes are also conspicuous SAT chromosomes, which have a prominent secondary constriction. Hosta seeds also have the typifying, sooty black coating around their seeds called phytomelanin, which is characteristic of the Asparagales. Thus, Hosta is clearly placed as Hostaceae in the higher Asparagales in which Agavaceae are close relatives. Yet, there is differentiating micro- and macro-data to confirm the placement of Hosta in the family Hostaceae in the order of Asparagales (Schmid 1991; Kubitz 2010). Further confirmation came from Kaneko (1968), who also found karyotypical differences between Hosta and Yucca-Agave to preclude even the probability of a common origin for them. D.A. Patil and R.M. Pai (2011) make the following statement: "Although Traub (1975) and Takhtajan (1980) place it (i.e. Hosta) in the Agavaceae, it is not apparently a comfortable placement. Observation of the stomatal complex would have given another parameter, which appears very pertinent as far as these taxa are concerned. A sum total of the evidence would seem to demonstrate that Hosta is a "half-way house" between the lilies and the agavoids but more at home with the latter than with the former." It is obvious by the data examined that placement in the monotypic Hostaceae is a much better taxonomic solution. This placement is also corroborated by Jia-Xi Liu et. al. (2011). Their detailed analysis concentrated on Hosta anther structure and pollen ontogeny to provide additional data for Hosta classification. Their applied data reveals reference values that effectively support the classification of Hosta in the Hostaceae.
   Other confirmation for Hostaceae placement is provided by (Markandeya 1978), who found that embryologically the LS and LP strands are distinct in Hosta and without the development of a commissural bundle as occurs in Yucca-Agave. A fusion of these strands would have resulted in a composite LS-LP bundle. Cheadle and Kosakai (1971) determined that the vessels in Hosta have scalariform perforation plates and that these are found only in roots. This is in contrast to the plates found in Agavaceae (Agave sensu lato), which are present in the roots as well as the leaves and are of the specialized simple perforate type. A further argument for the placement in Hostaceae is the distinct taxonomic and morphological isolation of the genus Hosta, which consists of herbaceous perennial species, while Yucca- Agave complex consists of arboreal (tree-like) species.
   Another important facet to consider in the placement is the evolution of Hosta. The genus Hosta is not native in North America. The latter is the birthplace of Agavaceae. Although many other Asian species crossed the Beringia land bridge between Siberia and Alaska some 12- to 15,000 years ago and there is no evidence that Hosta migrated. Schmid (1991) concludes that hostas must have evolved after the land bridge disappeared or they were for some reason not able to migrate to North America.
   The genus Hosta evolved before the last ice age in China, where the most primitive Hosta species, i.e. H. plantaginea, originated. From China, the propagules spread into Korea and throughout Japan (Schmid 1991). Different habitats led to increased adaptation and speciation. From an evolutionary standpoint, Hosta is a very young genus occurring in a contiguous area of Asia and is found nowhere else on Earth. It has unique macro- and micromorphology and no other ornamental plant looks like it. Its evolutionary habitat consists of subtropical, temperate, and subarctic areas where Hosta has locally adapted to complete its seasonal growth cycle ranging from 10 months in the subtropical regions (Kyushu, Japan) to 5 months in the subarctic (Hokkaido, Japan). Rainfall amounts average 70 inches (1780 mm) with additional East-Asian monsoon rains adding seasonal rainfall, which can double the total rain received in many areas of South-East Asia. In its northern habitat, heavy snowmelt supplies abundant moisture in spring. Hosta is acclimatized to shady forests of the mountains and the coast, in wetlands but also on well-drained rock outcrops. This has led to species differing in concert with their moisture requirements, which amounts to a minimum rainfall of 70 inches (1780 mm) (Schmid 1991).

Agave deserti (Agavaceae)
2005.03.28 in Anza-Borrego Desert State Park, California (©

   The above data place further doubt on the placement of Hosta with Agavaceae. Agave species are adapted to a xerophytic (desert) environment of semiarid to arid regions. The evolutionary center of origin is Mexico but local populations of Agavaceae families have spread from the southwestern U.S. through Central America, the Caribbean, and into northern South America. In most of the desert-like habitat total annual rainfall ranges from less than 5 inches to no more than 10 inches and for long periods there may be little or no precipitation. This desert area has a wide range of seasonal and daily temperatures. The mean maximum temperature in July is between 100°F and 105°F, and the minimum in January is 36° F. (Bowman, 1973)
   Recent phylogenetic analysis (S.V. Good-Avilas et al 2006) shows that the age of the species group of Agave (Agave sensu lato) is less than or equal to 10 million years. This is backed by fossil evidence. Hosta, on the other hand, shows an absence of fossil evidence. A diligent search of the scientific literature has not produced any indication that fossil evidence exists for Hosta (Schmid 1991). This leads to the hypothesis the genus must have evolved during the latter part of the Pleistocene period from 1.8 million to 8,000 years ago. Hosta further developed through speciation, inter-specific hybridization, and adaptation during the Holocene Period, from 8,000 years ago to the present time. Some species are probably still speciating and adapting to constantly changing environmental conditions.
   One of these speciation events was reported when morphometric multivariate analysis showed a very close relationship between H. minor and H. venusta (M.G. Chung, 1990). Based on macromorphology and distribution patterns Chung concluded that H. venusta (Cheju Island) is a recent derivative of H. minor (southeastern and southern peninsular Korea). Chung’s study shows that the propagules of H. venusta were moved from southeastern Korea to Cheju Island after the last glacial epoch. The geological age of Cheju Island is estimated to be ±13,000 years. H. venusta is found only  on this island and speciated after the volcanic island became habitable. During the process of adapting to a new, hostile, basaltic island habitat of recent origin, H. venusta underwent subsequent genetic changes that gave it the miniature size it is today. Chung was able to determine that 49 enzyme bands of H. venusta are a subset of 72 bands found in H. minor so establishing a very close relationship of the two species.
   Reports by Schmid (1991) and Yinger (2012; to be published in The Hosta Journal) describe another Hosta speciation event occurring after the last Ice Age and resulting in two distinct species named recently: H. yingeri (S.B. Jones (1989) and H. laevigata (W.G. Schmid 1991). These new species have a very distinct leaf and flower morphology. The perianth is spider-flowered with very narrow, purple lobes and bimodal anther lengths (3 long + 3 short). H. laevigata differs with its longer, narrow leaves with very undulate margins and a normal arrangement of anthers. Both have been placed in section Arachnanthae (Schmid 1991) reflecting the spider-flowered perianth unlike those of any other Hosta. The new species came into being after the last Ice Age (± 18,000 years), when rising sea levels isolated thousands of hills and mountains and turned them into islands in the Yellow Sea. The complete isolation of plants on remote islands and climatic changes resulted in the development of new species based on thousands of years of inbreeding. As reported in Schmid (1991) and Yinger (2012), inbreeding can lead to genetic drift, in which characteristics that are not necessarily adaptive can become fixed in the genome. These differentiating features found in the taxa of native Korean Hosta species indicate they evolved in isolation from the Japanese forms.


   Science has accepted the family name Hostaceae among the Angiosperms, the flowering plants we treasure so much. They have always been companions of our life cycles, be it to celebrate a new baby, our families birthdays, or to decorate the graves of our departed loved ones. The genus Hosta holds a special position among them. The flowering plants started their existence about 140 million years ago, long before the first hostas appeared. We do not know exactly when the first hostas evolved, but we can be certain that they are belonging to one of the youngest families among flowering plants, i.e. Hostaceae. They have a propensity for adapting to new and changing habitats and climates as the development of some of the Korean species shows. The members of the family Hostaceae will continue to surprise us with their ability to undergo genetic changes and we can be certain that more surprises wait for us in their native habitat.

H. ‘Sieboldiana’ (H. sieboldiana)
Munich Botanical Garden
Trial Garden of the Ludwig-Maxmilian University, Munich, Germany
© W.G. Schmid 1984.08.20

Read References              Go to Page 1

Reading Room
Library Homepage