Browsing by Subject "Erkennung"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Publication
    Ontogenetic and individual patterns of volatiles in honeybee queens Apis mellifera and its significance for the acceptance of queens in honeybee colonies
    (2008) Al Ali Alkattea, Raghdan; Bessei, Werner
    Activities of honeybees Apis mellifera L. colony are coordinated by an effective communication network in which the queen plays a central role by controlling behavior and reproduction of workers through pheromones. Most pheromones are produced in the mandibular (QMP) and tergal gland and distributed over the queen?s cuticle. The acquisition of these pheromones from the cuticular body surface of the queen is performed by antennating and licking of the retinue workers. Workers of a colony are able to recognize their own queen. Foreign queens which are introduced without protection are normally killed by the workers. While a lot of work has been performed on the primer and releaser effect of certain queen pheromones, it is still unknown how the workers distinguish their own queen from foreign ones. The fact that queens can be exchanged successfully by protecting the foreign queen for some days demonstrates that workers are able to ?learn? their queen. It is likely that a certain chemical pattern of the cuticle (odor or taste) is finally responsible for the recognition and acceptance as ?own?. In this context, this work has three different objectives: - To better understand the bees? behavior to ?own? and ?foreign? queens and to quantify certain behavioral traits of the queen-workers interaction. - To study the learning ability for own and foreign queens by the use of the Proboscis Extension Reflex (PER) in order to have a tool for future tests of odorous compounds. - To compare the cuticular pattern of queens of different origin. In all three approaches, virgin and mated queens and queens of different kin relation to each other were reared and established in Kirchhainer nuclei colonies. These queens were compared due to the following hypothesis: If the workers perceive their own queen by a distinct smell and if closely related queens have a more similar chemical pattern on the cuticle, then a related foreign queen should be easier ?learned?/ accepted than a non related one. For this purpose, first a specific bioassay had to be developed and established to enable the record of workers behavior to the queen without an inhibition of the complex social interactions between queens and workers. This ?cage bioassay? consists of a small wooden box with a glass front, a wax comb, 30-40 worker bees and a queen. For the tests, the own queen of this mini-colony was removed and a foreign queen was introduced. For a period of about 2 hours certain aggressive and benign actions, respectively, of the workers toward the queen were recorded. In the first set of tests, queens of different kin relations were compared. The results showed, in general, an aggressive reaction against the introduced foreign queens. However, there were clear lower benign and stronger aggression behaviors against unrelated queens compared to the related ones. Some of the unrelated queens were even killed. However, these differences were only significant when virgin queens were exchanged but not when mated queens were used. Concerning the duration of the aggressive action of workers, aggression generally decreased between the beginning and the end of the test; again, this was significant only in the experiments with virgin queens. This indicates, that at least in virgin queens the individual recognition by the worker bees depends on a kin specific odorous pattern of the queen. The same types of queens used in the cage bioassays were used for the learning experiments. A classical olfactory conditioning (PER) of worker honeybees was applied by using a living queen as the source of odor. Hereby the queen was offered in a way that the worker bees could not touch the body surface. The gradual increase in the learning curves was a good indication that the workers are able to learn the queen?s odor and, therefore, can be used as a kind of ?biosensor?. After having learned a queen?s odor, the conditioned workers were tested by offering virgin and mated queens, respectively, with defined kin relation to the queen used for the conditioning before. The results revealed clear differences in the cues used for the ?learning? of individual mated and individual virgin queens, respectively. The workers could significantly discriminate between the learned odor of a mated queen and any other mated queen irrespective of the relatedness. In contrast, worker bees could not discriminate virgin queens from each other. As the worker could only use volatile substances for the associative learning, one can conclude the following: In virgin queens the volatile ?bouquet? is neither individual specific nor kin specific. In mated queens the bouquet has only an individual specificity. Probably, the huge amount of many volatile gland products (including the main component 9-ODA) makes each mated queen ?unique?. But as in the cage bioassays the worker could recognize whether an introduced virgin queen was related to the own queen or not, these recognition must depend on non volatile substances of the virgin queens cuticle which are perceived by licking. If learned and tested queens were of different mating status the worker bees could significantly discriminate between such individuals (except learned odor of mated queens/ tested odor of related virgin queens). This is not surprising because the GC-MS analysis confirmed the huge differences in the odorous pattern (and here mainly the volatile polar gland products) between virgin (= young) and mated (=elder) queens. From the same types of queens used for the cage bioassays and PER, queens? heads and abdomens were extracted in a solvent and the obtained extracts were analyzed using GC-MS. From the extracts of queen abdomens 32 substances (hydrocarbons and polar compounds) were identified and chosen to calculate the ?chemical distance? between queens of different kin relation (sister vs. unrelated) and between sister queens having different ages and mating status. For that purpose, a matrix of Nei-distances was applied as a measure for the similarity of different patterns. The results showed a significantly higher concordance in the chemical pattern within sister queens compared to non related ones. The ?chemical distance? increased from sister queens over half sister to non-related queens. Cluster analyses of the Nei distance and multidimensional scaling clearly confirmed the differentiation between unrelated queens and the similarity of sister queens. Using the same statistical methods, also a clear differentiation between queens of different ages and mating status could be demonstrated. The results presented in this work confirmed with 3 different approaches that workers are able to learn their own queen with different learning cues depending on the mating status of the queen. In virgin queens it could be demonstrated for the first time that the kin relation between different queens can be recognized, presumably by low or non-volatile substances. The chemical analyses confirmed that the cuticular pattern of queens could be used for the differentiation not only according to age but also according to kin.
  • Thumbnail Image
    Publication
    Use of sensor technologies to estimate and assess the effect of various plant diseases on crop growth and development
    (2008) Gröll, Kerstin; Claupein, Wilhelm
    The topic of this study was ?Use of sensor technologies to estimate and assess the effect of various plant diseases on crop growth and development?. The background of the investigation can be seen in the challenge of developing a sensor system for the site-specific identification of plant diseases. The most widely used practice in disease control is still to spray fungicides uniformly over fields at different times during the vegetation period. However, most diseases are not distributed uniformly across a field, but occur in patches. During the early stage of epidemics large areas of the field are disease free. Excessive use of fungicides increases costs and can increase fungicides residue levels on agricultural products. As there is an increasing pressure to reduce their use by targeting fungicide spraying only on those places in the field where they are needed, the challenge is to provide farmers the the appropriate technological solutions. A simple and cost-effective optical device, based on the measurement of canopy reflectance in several wavebands, would allow disease patches to be identified and thus controlled. The implementation of these reflectance measurement data into crop growth models would allow for the development of site-specific decision rules whether to spray or not to spray. The specific objectives of the Ph.D. thesis were to: develop and test reflectance measurements as a possible technology to identify reflectance signatures of various plant diseases; develop suitable sets of calibrations that can be used for the identification and quantification of plant diseases; test different sensor systems at different spatial resolutions for their ability to identify plant diseases; develop a strategy to use plant disease information gained from sensor measurements as input dataset for the simulation of wheat growth under disease pressure in CERES-Wheat. In greenhouse experiments at the University of Hohenheim and in field experiments at the experimental station ?Ihinger Hof? of the University of Hohenheim the influence of the diseases powdery mildew, septoria leaf blotch and wheat eyespot on the reflectance of winter wheat was analyzed. To measure the reflectance of the plants three different sensor systems were used. Plant reflectance was measured with a digital camera (LEICA S1 PRO, LEICA Kamera AG, Solms, Germany) at leaf scale (0.5 cm²) and with the spectroradiometer Field Spec® Hand Held (ASD, Inc. Boulder, CO, USA) (0.5 m²) and the Yara N-Sensor in the field-scan modus (12 m²) 2 m above the canopy. The diseases powdery mildew, septoria leaf blotch and wheat eyespot have been analyzed. In a first approach it was tested if it is possible to detect plant diseases using reflectance measurements. The greenhouse studies showed that powdery mildew could be identified especially in the visible wavelength range. Also a correlation between powdery mildew pustules and reflectance changes was possible. Powdery mildew is a leaf disease and changes could directly be detected by a sensor system (Chapter 5). Out of this the second approach was to analyze if a stem disease that cannot directly be detected could be identified using a sensor system. The influence of wheat eyespot was investigated in a field experiment with winter wheat. The results showed that wheat eyespot could not be detected with the digital camera and the spectroradiometer. The problem was the low infection level and the distance between the measuring place and the infection place (Chapter 6). In a next step common vegetation indices were tested for their ability to identify plant diseases. Different vegetation indices were selected out of the literature to detect powdery mildew and septoria leaf blotch in the field using a spectroradiometer. Results indicated that the common vegetation index REIP was able to detect powdery mildew at an infection level of 7 %. With the common vegetation indices septoria leaf blotch could be detected only at a late infection level of 13.7 %. Out of this the new vegetation index DII was developed, which was able to detect septoria leaf blotch at an early infection level of 4 % (Chapter 7). Not only the place of infection but also the spatial resolution seems to play an important role in the identification of plant diseases. In a further approach different sensor systems with different spatial resolutions were tested in a field experiment for the identification of septoria leaf blotch. The results showed in general that septoria leaf blotch could be identified especially in the infrared wavelength range compared to powdery mildew that could especially identified in the visible wavelength range. The results showed further that the lower the spatial resolution , the more difficult it gets to identify plant diseases site-specifically. With a spatial resolution of 0.5 cm² a identification and quantification was possible. With a spatial resolution of 0.5 m² only a identification was possible and with a spatial resolution of 12 m² not identification and quantification was possible. That might be because of the resulting mixture of healthy and diseased plants (Chapter 8). The last step of this work was then to show how reflectance measurements could be implemented into crop growth models to calculate decisions whether to spray or not to spray fungicides on a site-specific level. Summarizing, the overall results of this study indicated that an identification of plant diseases was possible under certain conditions. An identification was possible if the infection place was also the measuring place and if a sensor system was used with a high spatial resolution. The results also showed that it was possible in a certain way to differ between biotroph and necrotroph plant diseases. For a holistic farming concept it is necessary in the future that reflectance measurements are integrated in a crop growth model to give farmers a decision tool that decides whether the infection is critical enough to spray or not.