Does monocarpic Phyllostachys nigra var. henonis regenerate after flowering in Japan? Insights from 3 years of observation after flowering

Abstract

Phyllostachys nigra var. henonis, a monocarpic bamboo with a 120-year flowering interval, is next predicted to flower in Japan in the 2020s. Because a huge area of the country is presently covered by stands of this species, post-flowering dieback of these stands and ensuing drastic changes in land cover may cause serious social and/or environmental problems. No study on the regeneration of this bamboo species was conducted during the last flowering event in the 1900s, and the regeneration process of this species is thus still unknown. In 2020, we encountered a localized flowering of P. nigra var. henonis in Japan and used this discovery as a rare opportunity to study the initial regeneration process of the species. Over 3 years, more than 80% of culms in the study site bloomed, but no seed was produced. In addition, no established seedlings were located. These facts strongly suggest that P. nigra var. henonis lacks the ability to produce seeds and cannot undergo sexual regeneration. Some bamboo culms were produced after flowering but died within 1 year of emergence. Small, weak culms (dwarf ramets) also appeared after flowering, but most died within 1 year as well. Three years after flowering, all culms had died, with no sign of regeneration detected. According to our 3 years of observation, this bamboo appears to be hard to regenerate-an idea completely contradicted by the fact that this species has long persisted in Japan. We thus considered other possible regeneration modes for P. nigra var. henonis.

Fig 1. Flowering rates of Phyllostachys nigra var. henonis in 4 × 4 m² subquadrats in June 2020 (A), June 2021 (B), June 2022 (C), and over 3 years, from 2020 to 2022 (D), in Fukutomi, Higashi-Hiroshima, Japan.

Rates are the number of flowering culms divided by the initial number of culms in each SQ multiplied by 100. The locations of seed traps are also shown in Fig 1A (●).

Fig 2. Frequency distribution of the dbh of flowering and non-flowering culms of P. nigra var. henonis in Fukutomi, Higashi-Hiroshima, Japan, over 3 years, from 2020 to 2022.

Flowering culms were significantly thicker than non-flowering ones (Wilcoxon rank sum test with continuity correction, W = 14,370, P < 0.001).

Fig 3. Spatial distribution of 19 normal-sized culms asexually produced by P. nigra var. henonis after flowering in 2020 in Fukutomi, Higashi-Hiroshima, Japan.

The flowering rates of SQs are also shown. Normal-sized culms emerged only in SQs with low flowering rates. Nineteen and zero culms were produced in 2021 and 2022, respectively.

Fig 4. Frequency distribution of the height of dwarf ramets of P. nigra var. henonis in Fukutomi, Higashi-Hiroshima, Japan.

The fates of the ramets 1 year after emergence are also shown. Dwarf ramets were dying after flowering. At least 415 dwarf ramets were produced, but only 75 survived for 1 year.

Fig 5. Survivorship of flowering and non-flowering culms of P. nigra var. henonis from June 2020 to October 2022 in Fukutomi, Higashi-Hiroshima, Japan.

Our first census was taken in January 2021. Hence, the number of culms in June 2020 was estimated based on the results the first census; number of non-flowering culms in June 2020 was estimated as the same number of non-flowering culms observed in January 2021 (shown by an open circle). Consequently, survivorship between June 2020 and January 2021 (denoted by a broken line) may be overestimated, as we assumed that no non-flowering culms died during this period. Most flowering culms died within 1 year of flowering, and almost all culms (95%) died by June 2022 (2 years after flowering). In contrast, non-flowering culms rarely died over the same period (mortality rate of 5%). Regardless of flowering status, all culms in the plot died by October 2022.

Fig 6. Spatial pattern of canopy openness (%) in April 2021 (A) and relationship between flowering ratio in 2020 (%) and canopy openness (%) (B).

Almost all culms flowering in June 2020 had already died and shed their leaves when canopy openness was measured in April 2021; consequently, canopy openness was quite high in SQs that had many flowering culms in 2020 (see Fig 1A). A thick line shown on (B) represents a regression line; Y = 0.167X + 7.65, R² = 0.65.

Fig 7. Spatial pattern of ground vegetation coverage in July 2021, when annual herbs and grasses were fully grown (A) and relationship between canopy openness (%) and ground covered by vegetation (%) (B).

Ground vegetation grew well in SQs having a high canopy openness (see Fig 6A). Ground covered by vegetation was regressed with canopy openness by using a logistic function: Y = 100/(1 + 207 e^-0.332X), R² = 0.892 for (a thick line on (B)).