Mirror‐image flowers represent a highly specialized pollination system that is normally associated with poricidal anthers and buzz pollination (pollen can be only released from anthers via pollinator vibration). In the Asian endemic Hiptage benghalensis (Malpighiaceae), mirror‐image flowers were found with longitudinally dehiscing anthers and such floral syndromes are firstly reported in the family. We investigated the floral biology and frequencies of left‐ and right‐styled flowers of H. benghalensis in Hainan Island, southern China. We conducted pollination manipulations to determine breeding systems and studied the pollination mechanism. We also examined the compositions of the secretions of the calyx gland. Controlled pollinations revealed that H. benghalensis is self‐compatible. Left‐ and right‐styled flowers are produced in the same inflorescence and a 1:1 ratio is found at both plant and population levels. Flowers have extremely reflexed petals, zygomorphic corollas and a single but oversized calyx gland. This gland secretes sugars and mainly attracts ants and wasps in both blooming and fruiting periods. Heteranthery is obvious with one large stamen producing more viable pollen than the nine small stamens. All anthers dehisce longitudinally and pollen grains readily adhere to floral visitors without the need of buzz pollination. The unusual association of mirror‐image flowers with longitudinal anthers probably reflects an adaptation to local pollen‐collecting (but non‐buzzing) honeybees such as Apis dorsata. These results indicate that floral syndromes and pollination adaptation in H. benghalensis differ completely from the New World Malpighiaceae and may help to explain evolutionary adaptations of the family during its long‐distance dispersals from the New World to Asia. © 2013 The Linnean Society of London, Botanical Journal of the Linnean Society, 2013, 173, 764–774.
Aims Modelling potential distribution ranges of threatened species is of great significance for their conservation. In this paper, the distribution of potential suitable habitat of Impatiens hainanensis, a limestone-endemic and endangered plant in Hainan Island, was studied to provide scientific basis for their effective in situ conservation and re-introduction of I. hainanensis. Methods Based on eight occurrence sites and 12 environmental variables, the Maximum Entropy (MaxEnt) algorithm and GIS technology were used to create a model linking the distribution ranges of I. hainanensis with environments. With data on five actual distribution sites and five non-occurrence sites, four model evaluation metrics (area under the receiver operating characteristic curve (AUC), kappa coefficient, true skill statistic (TSS), overall accuracy) were used to evaluate the predictive performance and accuracy of this model. Important findings The results indicated that the indicative value of all four evaluation metrics were above 0.9, indicating that the MaxEnt model can effectively predict the potential suitable habitats of I. hainanensis. Slope, precipitation of the driest quarter and coefficients of precipitation variation were the three main environmental factors influencing the distribution of I. hainanensis. At present, the most suitable habitat includes western and southern parts of Baisha County, the central and southern parts of Changjiang County, the eastern part of the Dongfang City and northeastern Ledong County, accounting for 1.8% of land area on Hainan Island. Since the potential suitable habitat of I. hainanensis is rare and severely fragmented, the protection of this species is urgent. We suggest to collect the seeds of various geographic populations of I. hainanensis to establish a germplasm resource bank. The most suitable habitat of the species, including Tian'an Village and Jiangbian Village in
The southwestern mountains of Hainan Island are the southernmost region with tropical karst landform in China. The frequent alternation of dry and wet seasons leads to the loss of the mineral nutrients of limestone, creating karst fissure habitats. Plants living in karst fissure habitats for long periods of time have developed local adaptation mechanisms correspondingly. In the paper, hydrogen–oxygen stable isotope technology was applied to determine the water-use sources of Impatiens hainanensis in the dry and wet seasons, hoping to expound the adaptation mechanism of I. hainanensis in karst fissure habitats to the moisture dynamics in the wet and dry seasons. In the wet season (May to October, 2018), the air humidity is relatively high in the I. hainanensis habitat; in the dry season (November 2018 to April 2019), there is a degree of evaporation. In the wet season, fine-root biomass increases with soil depths, while coarse-root biomass decreases with soil depths; in the dry season, fine-root biomass is lower and coarse-root biomass is higher compared with the wet season. It was found that the average rainfall reached 1523 mm and the main water-use sources were shallow (0–5 cm) and middle (5–10 cm) soil water, epikarst water, and shallow karst fissure water during the wet season; the average rainfall reached 528 mm, and the deep (10–15 cm) soil water and shallow karst fissure water were the main water-use sources during the dry season. Fog water has a partial complementary effect in the dry season. The differences in the distribution of root biomass and each source of water in the wet and dry seasons of I. hainanensis also reflect the different water-use strategies of I. hainanensis in the wet and dry seasons. In both dry and wet seasons, I. hainanensis formed a water-use pattern dominated by soil water and shallow fissure water (0–15 cm) under the influence of the “fissure-soil-plant” system in the karst region.
The accurate classification of forest types is critical for sustainable forest management. In this study, a novel multiscale global graph convolutional neural network (MSG-GCN) was compared with random forest (RF), U-Net, and U-Net++ models in terms of the classification of natural mixed forest (NMX), natural broadleaved forest (NBL), and conifer plantation (CP) using very high-resolution aerial photographs from the University of Tokyo Chiba Forest in central Japan. Our MSG-GCN architecture is novel in the following respects: The convolutional kernel scale of the encoder is unlike those of other models; local attention replaces the conventional U-Net++ skip connection; a multiscale graph convolutional neural block is embedded into the end layer of the encoder module; and various decoding layers are spliced to preserve high- and low-level feature information and to improve the decision capacity for boundary cells. The MSG-GCN achieved higher classification accuracy than other state-of-the-art (SOTA) methods. The classification accuracy in terms of NMX was lower compared with NBL and CP. The RF method produced severe salt-and-pepper noise. The U-Net and U-Net++ methods frequently produced error patches and the edges between different forest types were rough and blurred. In contrast, the MSG-GCN method had fewer misclassification patches and showed clear edges between different forest types. Most areas misclassified by MSG-GCN were on edges, while misclassification patches were randomly distributed in internal areas for U-Net and U-Net++. We made full use of artificial intelligence and very high-resolution remote sensing data to create accurate maps to aid forest management and facilitate efficient and accurate forest resource inventory taking in Japan.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.