To solve the issue of insufficient pollinating of insects for the oil tree peony ‘Feng Dan’ (Paeonia ostii T. Hong et J.X. Zhang) and improve its seed set and yield, we conducted observations from 2017 to 2018 to investigate the relationship between honey bee (Apis mellifera L.) foraging behavior and diurnal activity. We compared the single-fruit seed set ratio among three flower types on the same plants of the oil tree peony, which flowered simultaneously, in three pollination areas (bee pollination, natural field pollination, and controlled pollination by pollinators) and in a net room under self-pollination, wind pollination and bee pollination. Apis mellifera exhibited short single visitations, long visitations to a single flower and repeated visits to flowers of the oil tree peony. The number of flower visits of A. mellifera was significantly and positively yet weakly correlated with the number of stigma visits (2017: r = 0.045, p < 0.05; 2018: r = 0.195, p < 0.01). The seed set of oil tree peony follicles in the A. mellifera pollination area was significantly higher than that in the natural pollination field area and the control net rooms. On the same oil tree peony plant with synchronous flowering, the percent seed set of follicles pollinated by A. mellifera at a high density was significantly higher than that resulting from wind pollination and self-pollination.

AbstractOil peony (Paeonia spp.) is a new type of woody oil crop in China with a large cultivation area. Inadequate pollination is one of the main reasons for low seed yield. A pollination net room was built at an oil tree peony base, the numbers of honeybees (Apis mellifera L.) and ground bumblebees (Bombus terrestris L.) were artificially increased and the foraging behaviors and daily activities of the two bees on the plants were observed. Four different pollination methods (honeybee pollination, ground bumblebee pollination, natural field pollination and pollination without insects) were applied. The visit interval, single-flower visit time, number of single-flower visits, number of flowers visited per minute and number of stigma contacts were compared. Meanwhile, the effects of honeybee and bumblebee pollination on the oil yield and quality of peony seeds were compared. There were noticeable differences in daily activity between honeybees and ground bumblebees. Significant differences in the single-flower visit time, visit interval and visit frequency were also observed; honeybee and ground bumblebee pollination increased the seed yield of oil tree peony by 78.74% and 31.88%, respectively. Therefore, both honeybees and ground bumblebees are effective pollinators of oil tree peony. These results provide a theoretical basis for further utilization of bee resources for oil tree peony pollination.

The diversity of arbuscular mycorrhizal fungi (AMF) is of great interest because of their potential function in ecosystems. Tree peony is an important traditional ornamental and medicinal plant with economic significance. We examined the mycorrhizal status of the rhizosphere of 14 common cultivars of tree peony (Paeonia suffruticosa) in 3 different geographic locations in China. Root samples of all cultivars were colonized by AMF. The mean percentage of root length colonization, vesicles, and arbuscules were 39%, 3.6%, and 6.0%, respectively. AMF species richness varied from 5 to 11, and spore density ranged from 20 to 61 per 50 g of rhizospheric soil. The average AMF species diversity (Shannon-Wiener index) was 1.92, ranging from 1.64 to 2.18. A total of 31 AMF species belonging to 3 genera were identified in the rhizospheric soil. Glomus (21) was the dominant genus, followed by Acaulospora (7) and Scutellospora (3). G. aggregatum was the most commonly distributed species, with an occurrence frequency of 71.4 and a relative abundance of 13.6%. This study focused on the comparison of AM fungal diversity associated with tree peony in various original cultivar groups. This knowledge will help in selecting suitable AM inoculums for cultivation in the different original cultivar groups of tree peony.

Background The study of plant phenotype by deep learning has received increased interest in recent years, which impressive progress has been made in the fields of plant breeding. Deep learning extremely relies on a large amount of training data to extract and recognize target features in the field of plant phenotype classification and recognition tasks. However, for some flower cultivars identification tasks with a huge number of cultivars, it is difficult for traditional deep learning methods to achieve better recognition results with limited sample data. Thus, a method based on metric learning for flower cultivars identification is proposed to solve this problem. Results We added center loss to the classification network to make inter-class samples disperse and intra-class samples compact, the script of ResNet18, ResNet50, and DenseNet121 were used for feature extraction. To evaluate the effectiveness of the proposed method, a public dataset Oxford 102 Flowers dataset and two novel datasets constructed by us are chosen. For the method of joint supervision of center loss and L2-softmax loss, the test accuracy rate is 91.88%, 97.34%, and 99.82% across three datasets, respectively. Feature distribution observed by T-distributed stochastic neighbor embedding (T-SNE) verifies the effectiveness of the method presented above. Conclusions An efficient metric learning method has been described for flower cultivars identification task, which not only provides high recognition rates but also makes the feature extracted from the recognition network interpretable. This study demonstrated that the proposed method provides new ideas for the application of a small amount of data in the field of identification, and has important reference significance for the flower cultivars identification research.

No abstract