Information is lacking regarding the visual cues used by Helicoverpa armigera moths during nectar feeding. Here, we investigated the preference for radial gradient patterns in H. armigera moths. The results indicated ...Information is lacking regarding the visual cues used by Helicoverpa armigera moths during nectar feeding. Here, we investigated the preference for radial gradient patterns in H. armigera moths. The results indicated that both sexes shared a preference to plain flower models of blue and cyan. The radial gradient pattern (cyan as nectar guide color and blue as petal color) was more attractive than its component plain colors (cyan or blue), especially in male moths. Number of corolla petals did not influence the attractiveness of the cyan-blue pattern. The addition of a tertiary floral attractant to white-blue or cyan-blue radial gradient patterns could dramatically enhance the attractiveness of visual cues in males rather than females, suggesting that males gave a higher weight in olfactory modality than females gave, while females gave a higher weight in vision modality than males gave. All together, we found an optimal combination of floral cues in H. armigera sexes as follows: A tertiary floral attractant (2 μL dose of phenylacetaldehyde, benzyl acetate, and salicylaldehyde mixed in 26:15:2) added to white-blue radial gradient flower model (3 cm in diameter). To our knowledge, this is the first time that rose curve and radial gradient tools were used to simulate floral pattern in the studies of flower-visiting insects.展开更多
Estimation of the transpiration rate for a tree is generally based on sap flow measurements within the hydro-active stem xylem. In this study, radial variation of sap flow velocity（Js） was investigated at five depth...Estimation of the transpiration rate for a tree is generally based on sap flow measurements within the hydro-active stem xylem. In this study, radial variation of sap flow velocity（Js） was investigated at five depths of the xylem（1, 2, 3, 5 and 8 cm under the cambium） in three mature Xinjiang poplar（Populus alba L. var. pyramidalis） trees grown at the Gansu Minqin National Studies Station for Desert Steppe Ecosystem from May to October 2011. Thermal dissipation probes of various lengths manufactured according to the Granier＇s design were installed into each tree for simultaneous observation of the radial patterns of Js through the xylem. The radial patterns were found to fit the four-parameter GaussAmp equation. The peak Js was about 27.02±0.95 kg/（dm2？d） at approximately 3 to 5 cm deep from the cambium of the three trees,and the lowest Js appeared at 1 cm deep in most of the time. Approximately 50% of the total sap flow in Xinjiang poplar occurred within one-third of the xylem from its outer radius, whereas 90% of the total sap flow occurred within two-fifth of the xylem. In addition, the innermost point of the xylem（at 8-cm depth）, which appeared as the penultimate sap flow in most cases during the study period, was hydro-active with Js,8 of 7.55±3.83 kg/（dm2？d）. The radial pattern of Js was found to be steeper in midday than in other time of the day, and steeper diurnal fluctuations were recorded in June, July and August（the mid-growing season）. Maximum differences between the lowest Js（Js,1 or Js,8） and the highest Js（Js,3 or Js,5） from May through October were 12.41, 17.35, 16.30, 18.52, 12.60 and 16.04 g/（cm2？h）, respectively. The time-dependent changes of Js along the radial profile（except at 1-cm depth） were strongly related to the reference evapotranspiration（ET0）. Due to significant radial variability of Js, the mean daily sap flow at the whole-tree level could be over-estimated by up to 29.69% when only a single probe at depth of 2 cm was used. However, the accuracy of the estimation of sap flow in Xinjiang poplar could be significantly improved using a correction coefficient of 0.885.展开更多
In this paper, an incremental learning model called Resource Allocating Network with Long-Term Memory (RAN-LTM) is extended such that the learning is conducted with some autonomy for the following functions: 1) data c...In this paper, an incremental learning model called Resource Allocating Network with Long-Term Memory (RAN-LTM) is extended such that the learning is conducted with some autonomy for the following functions: 1) data collection for initial learning, 2) data normalization, 3) addition of radial basis functions (RBFs), and 4) determination of RBF cen-ters and widths. The proposed learning algorithm called Autonomous Learning algorithm for Resource Allocating Network (AL-RAN) is divided into the two learning phases: initial learning phase and incremental learning phase. And the former is further divided into the autonomous data collection and the initial network learning. In the initial learning phase, training data are first collected until the class separability is converged or has a significant dif-ference between normalized and unnormalized data. Then, an initial structure of AL-RAN is autonomously determined by selecting a moderate number of RBF centers from the collected data and by defining as large RBF widths as possible within a proper range. After the initial learning, the incremental learning of AL-RAN is conducted in a sequential way whenever a new training data is given. In the experiments, we evaluate AL-RAN using five benchmark data sets. From the experimental results, we confirm that the above autonomous functions work well and the efficiency in terms of network structure and learning time is improved without sacrificing the recognition accuracy as compared with the previous version of AL-RAN.展开更多
文摘Information is lacking regarding the visual cues used by Helicoverpa armigera moths during nectar feeding. Here, we investigated the preference for radial gradient patterns in H. armigera moths. The results indicated that both sexes shared a preference to plain flower models of blue and cyan. The radial gradient pattern (cyan as nectar guide color and blue as petal color) was more attractive than its component plain colors (cyan or blue), especially in male moths. Number of corolla petals did not influence the attractiveness of the cyan-blue pattern. The addition of a tertiary floral attractant to white-blue or cyan-blue radial gradient patterns could dramatically enhance the attractiveness of visual cues in males rather than females, suggesting that males gave a higher weight in olfactory modality than females gave, while females gave a higher weight in vision modality than males gave. All together, we found an optimal combination of floral cues in H. armigera sexes as follows: A tertiary floral attractant (2 μL dose of phenylacetaldehyde, benzyl acetate, and salicylaldehyde mixed in 26:15:2) added to white-blue radial gradient flower model (3 cm in diameter). To our knowledge, this is the first time that rose curve and radial gradient tools were used to simulate floral pattern in the studies of flower-visiting insects.
基金supported by the National Natural Science Foundation of China (31070628)Field support for this research was provided by Gansu Minqin National Studies Station for Desert Steppe Ecosystem
文摘Estimation of the transpiration rate for a tree is generally based on sap flow measurements within the hydro-active stem xylem. In this study, radial variation of sap flow velocity（Js） was investigated at five depths of the xylem（1, 2, 3, 5 and 8 cm under the cambium） in three mature Xinjiang poplar（Populus alba L. var. pyramidalis） trees grown at the Gansu Minqin National Studies Station for Desert Steppe Ecosystem from May to October 2011. Thermal dissipation probes of various lengths manufactured according to the Granier＇s design were installed into each tree for simultaneous observation of the radial patterns of Js through the xylem. The radial patterns were found to fit the four-parameter GaussAmp equation. The peak Js was about 27.02±0.95 kg/（dm2？d） at approximately 3 to 5 cm deep from the cambium of the three trees,and the lowest Js appeared at 1 cm deep in most of the time. Approximately 50% of the total sap flow in Xinjiang poplar occurred within one-third of the xylem from its outer radius, whereas 90% of the total sap flow occurred within two-fifth of the xylem. In addition, the innermost point of the xylem（at 8-cm depth）, which appeared as the penultimate sap flow in most cases during the study period, was hydro-active with Js,8 of 7.55±3.83 kg/（dm2？d）. The radial pattern of Js was found to be steeper in midday than in other time of the day, and steeper diurnal fluctuations were recorded in June, July and August（the mid-growing season）. Maximum differences between the lowest Js（Js,1 or Js,8） and the highest Js（Js,3 or Js,5） from May through October were 12.41, 17.35, 16.30, 18.52, 12.60 and 16.04 g/（cm2？h）, respectively. The time-dependent changes of Js along the radial profile（except at 1-cm depth） were strongly related to the reference evapotranspiration（ET0）. Due to significant radial variability of Js, the mean daily sap flow at the whole-tree level could be over-estimated by up to 29.69% when only a single probe at depth of 2 cm was used. However, the accuracy of the estimation of sap flow in Xinjiang poplar could be significantly improved using a correction coefficient of 0.885.
文摘In this paper, an incremental learning model called Resource Allocating Network with Long-Term Memory (RAN-LTM) is extended such that the learning is conducted with some autonomy for the following functions: 1) data collection for initial learning, 2) data normalization, 3) addition of radial basis functions (RBFs), and 4) determination of RBF cen-ters and widths. The proposed learning algorithm called Autonomous Learning algorithm for Resource Allocating Network (AL-RAN) is divided into the two learning phases: initial learning phase and incremental learning phase. And the former is further divided into the autonomous data collection and the initial network learning. In the initial learning phase, training data are first collected until the class separability is converged or has a significant dif-ference between normalized and unnormalized data. Then, an initial structure of AL-RAN is autonomously determined by selecting a moderate number of RBF centers from the collected data and by defining as large RBF widths as possible within a proper range. After the initial learning, the incremental learning of AL-RAN is conducted in a sequential way whenever a new training data is given. In the experiments, we evaluate AL-RAN using five benchmark data sets. From the experimental results, we confirm that the above autonomous functions work well and the efficiency in terms of network structure and learning time is improved without sacrificing the recognition accuracy as compared with the previous version of AL-RAN.
