Species of Chamaelirium are known to be highly diversified in sexual characters, but there has been no detailed empirical approach to elucidate the mechanism and process of sexual diversification in the genus. This paper reports the results of population-based analyses on various sexual aspects of C. hisauchianum endemic to Japan. Of the 20 populations surveyed, eight (40%) were composed of facultative hermaphrodites highly labile in sexual expression, exhibiting a wide range of gradational variation from hermaphroditism through andromonoecy to male. The remaining 12 populations (60 %) consisted of females and facultative hermaphrodites. In them, 11.8–30.9% (mean 22.2%) of the individuals were female. With no reciprocal sex changes, females and facultative hermaphrodites were regarded as, plausibly genetically determined, discrete sex morphs. In contrast, sexual expression in both sex morphs appeared resource dependent. Besides the difference in floral sexual traits, the average ratio of ovuliferous flowers in females was higher than that in facultative hermaphrodites. The production of ovuliferous flowers by females was less susceptible to both field and cultivated conditions and tended to be significantly relatively higher under unfavorable growing conditions. Females were thus less plastic (or more persistent) in producing ovuliferous flowers (or female organs) than facultative hermaphrodites irrespective of internal and external conditions. In all populations, andromonoecious plants dominated (50.8–100%), which appeared to reflect the high plasticity in sex allocation of facultative hermaphrodites. The different modes of sex allocation were regarded as secondary sex characters and presumably stem from a difference in easiness or attainability of sexual reproduction between the two sex morphs. 

Survey of sex morphs in populations

A total of 20 populations of Chamaelirium hisauchianum, 18 are of subsp. kurohimense, and one each of subsp. hisauchianum and subsp. minoense, were surveyed in yearly flowering seasons mainly in the 1980s (Table 1). When populations were comparatively small, scapes (flowering stems) of all flowering plants were harvested and their sexual characters were examined in detail at a laboratory of Teikyo University at Hachioji, Tokyo, Japan. When populations were large, two to eight patches with an area from 1 m × 1 m to 7 m × 11 m were selected, and scapes of all flowering plants occurring therein were harvested and analyzed. The sex morph of a spike or plant was directly determined by calculating ratios of the component flower sex morphs (including abortive asexual flowers if present), or indirectly assessed by categorizing estimated ratios of component flower morphs into one of the morphs defined below (e.g. F1, AM2, M3). Though rare, when a plant has two or more flowering scapes, the ratios of flower morphs were averaged and the mean was taken to represent the plant. 

Observation of floral characters and measurements of floral parts

Observations of floral characters and measurements of floral parts were made under a stereomicroscope. Lengths of well-developed tepals (upper tepals of flowers) and filaments were measured for at least three representative flowers selected from each spike. The measurements were averaged and the mean value for each spike was used. Where applicable, Student’s t-tests were used to statistically test the significance of differences between two groups of samples. 

Survey of changes in sex form or expression under cultivation

Plants sampled from natural habitats were transplanted into the experimental botanical nursery of Teikyo University. Each plant was planted individually with an accession number in a pot. Using them, changes in spike length, total flower number (including abortive asexual flowers) per spike, and the combination and/or ratio of flower sex morphs per spike were successively monitored and recorded in each flowering season for up to seven consecutive years (i.e. up to eight flowering seasons, including the initial one when plants were first transplanted from natural habitats). Data obtained during cultivation were pooled and integrally compared with data obtained from natural habitats just prior to cultivation.

List of abbreviations for sexual characters

• A: Abortive and asexual.
  • A(-H): Innately hermaphroditic plant (spike) bearing only abortive flowers.
  • A(-F): Innately female plant (spike) bearing only abortive flowers.
• AM: Andromonoecious.
  • AM1: Andromonoecious plant (spike) with a ratio of hermaphroditic flowers > ca. 2/3 of all flowers borne (abortive flowers ignored).
  • AM2: Andromonoecious plant (spike) with a ratio of hermaphroditic flowers ca. 1/3–2/3 of all flowers borne (abortive flowers ignored).
  • AM3: Andromonoecious plant (spike) with a ratio of hermaphroditic flowers < ca. 1/3 of all flowers borne (abortive flowers ignored).
• F: Female.
  • F1: Female plant (spike) with a ratio of female flowers > ca. 2/3 of all flowers borne (including abortive flowers).
  • F2: Female plant (spike) with a ratio of female flowers ca. 1/3–2/3 of all flowers borne (including abortive flowers).
  • F3: Female plant (spike) with a ratio of female flowers < ca. 1/3 of all flowers borne (including abortive flowers).
• H: Hermaphrodite.
  • Fac. H: Facultative hermaphrodites, conditionally exhibiting hermaphroditism, andromonoecy, male or abortive/asexual state.
• M: Male
  • M1: Male plant (spike) with a ratio of male flowers > ca. 2/3 of all flowers borne (including abortive flowers).
  • M2: Male plant (spike) with a ratio of male flowers ca. 1/3–2/3 of all flowers borne (including abortive flowers).
  • M3: Male plant (spike) with a ratio of male flowers < ca. 1/3 of all flowers borne (including abortive flowers).