【作者】 曹振军；

【导师】 李保平；

【作者基本信息】 南京农业大学 ， 农业昆虫与害虫防治， 2010， 博士

【摘要】 从原产地引进天敌是一条有效防治外来入侵杂草的途径,然而如何评估其对非靶标植物的潜在风险,从而减少新的生物入侵问题,是传统生防实践和理论研究面临的严峻课题。广聚萤叶甲原产于北美洲,传入我国后主要取食外来入侵豚草,是防治豚草的潜在天敌。由于其并非严格的单食性昆虫,故对入侵地非靶标植物的潜在风险值得关注。本论文将寄主专一性测定的经典方法与新近提出的概念和方法相结合,将室内与室外、选择性与非选择性实验相结合,对广聚萤叶甲的产卵选择专一性进行了较为全面的研究,为评估和预测其潜在风险提供了丰富的依据,并为杂草生物防治实践的改进进行了有益的探索。1.室内条件下广聚萤叶甲的产卵选择引用识别期概念及Cox比例风险模型分析方法,结合经典观测方法,比较了靶标植物(豚草)与近缘非靶标植物(三裂叶豚草、苍耳、菊芋、食用向日葵品种和油用向日葵品种)在不同取食经历和不同日龄条件下,被广聚萤叶甲成虫选择产卵的风险。结果表明,广聚萤叶甲在靶标豚草与其它近缘非靶标植物之间,表现出对豚草很强的偏好；不同取食经历和不同日龄不显著影响广聚萤叶甲对靶标植物(豚草)的选择频次,对识别期也无影响。广聚萤叶甲在室内研究条件下具有很强的寄主专一性。2.野外受控条件下广聚萤叶甲的产卵选择鉴于行为因素不同程度地影响昆虫的寄主选择表现,分别在野外大型罩笼内和野外开放条件下,观察了广聚萤叶甲对靶标植物和非靶标植物的产卵选择行为。研究表明：(i)当有靶标豚草存在的情况下,广聚萤叶甲成虫明显偏好在豚草上产卵,豚草被选择产卵的风险明显高于其他测试植物；(ii)当没有靶标豚草存在的情况下,广聚萤叶甲成虫首先选择三裂叶豚草和苍耳产卵,在食用向日葵和油用向日葵上产卵的风险很低。广聚萤叶甲在野外条件下所表现出与室内研究中同样强的寄主专一性。3.野外开放条件下广聚萤叶甲的产卵选择本研究就广聚萤叶甲在我国南京地区不同生境下的野外产卵选择进行了调查。在豚草生长季节定期定点调查以下因素：株高、植株斑块大小、植株受光与否、植株开花与否、植株被害程度等,用广义线性混合模型对所调查的变量进行拟合,从而确定影响广聚萤叶甲产卵选择的主要因子。结果表明,斑块大小对广聚萤叶甲的产卵选择有明显的影响,但光照没有明显的影响。被害等级和开花随季节动态对广聚萤叶甲的产卵选择有明显的影响,而株高随季节动态没有明显的影响。在两因子互作影响上,除斑块分别与被害等级和株高互作没有明显的互作外。其他各因子之间的双因素互作均对广聚萤叶甲产卵有明显影响。4.广聚萤叶甲对不同植物的产卵选择策略寄主品质(如碳、氮和防御代谢物)影响昆虫的产卵策略一权衡卵数量、大小和品质。本试验通过研究广聚萤叶甲在不同植物对的产卵识别期、窝卵数大小和卵粒大小等变量,分析其产卵策略。研究结果表明,广聚萤叶甲明显偏好豚草,在豚草上的窝卵数大(平均值±标准差=20.3±1.5)而且卵粒亦大(0.077±0.001 mm3),二者无相关性；在非靶标的苍耳(10.7±1.1和0.059±0.001 mm3)和向日葵(10.2±0.9和0.063±0.001 mm3)上的窝卵数和卵粒大小都小于豚草,而且窝卵数与卵粒大小之间存在显著的负线性关系,即采取了“多而小”或“少而大”的产卵策略。5.广聚萤叶甲对豚草的产卵选择策略寄主植物密度和品质均可能影响昆虫的产卵策略。试验研究了广聚萤叶甲对不同大小豚草斑块的产卵选择、逐次(日)产卵行为以及在不同品质和分别感染不同比例卵块和幼虫的豚草上的产卵策略。研究结果表明,豚草品质影响广聚萤叶甲的产卵策略,但豚草密度没有影响。产卵前期尽可能的产多而大的卵粒；中期窝卵数和卵粒大小虽都略有下降,但存在明显的线性关系；产卵后期窝卵数和卵粒大小继续下降,两者没有明显的线性关系。广聚萤叶甲成虫明显偏好生长期豚草,且差异显著。叶甲在处于生长期的豚草上尽可能多的繁殖多(20.3±1.5)而大(0.077±0.001mm3)的后代；在苗期(14.7±1.3和0.074±0.001 mm3)和开花期(10.7±0.8和0.073±0.001 mm3)的豚草上的窝卵数和卵粒大小都小于在生长期豚草上的表现,且窝卵数与卵粒大小存在明显的负线性关系。广聚萤叶甲在感染不同程度卵块和幼虫豚草上的产卵结果表明：(i)广聚萤叶甲在着卵块比例较低(0%、25%、50%)的豚草上产的窝卵数(>15粒),显著大于着卵比例较高(70%、100%)的处理豚草(≈10粒),而且窝卵数随豚草上着卵块比例的增大而减小；(ii)广聚萤叶甲在2个中等程度着卵块比例(25%、50%)处理豚草上产的卵粒大小相近,均显著小于无卵块和较高卵块比例(75%、100%)处理豚草上的卵粒,广聚萤叶甲产出的卵粒大小与豚草已有的着卵块比例之间存在“U”型曲线关系：(iii)产卵风险在没有感染幼虫和感染程度为25%的豚草之间没有显著差异,但与其他感染幼虫处理的豚草存在显著差异；(ⅳ)广聚萤叶甲在感染幼虫较多的处理豚草(50%～100%)上产的窝卵数均较小(<11.2粒),显著小于在感染幼虫较少(25%)和对照豚草上的窝卵数(>21.6粒),广聚萤叶甲的窝卵数大小与豚草幼虫感染程度之间存在显著负相关线性关系；(ⅴ)广聚萤叶甲产在感染幼虫较少(25%)的处理豚草上的卵粒(0.068±0.001mm3)显著小于其他处理豚草(>0.077 mm3),在其他处理豚草之间无显著差异(图6-10),卵粒大小与豚草上已有的幼虫比例之间存在显著的正相关线性关系。上述若干实验说明,豚草上已有卵块和幼虫均影响广聚萤叶甲的产卵选择以及窝卵数与卵粒大小的权衡策略。6.广聚萤叶甲成虫和初孵幼虫的取食选择寄主植物缺乏会迫使植食性昆虫的寄主选择发生变化,从而影响其寄主专一性。本实验首先对广聚萤叶甲成虫的耐饥饿能力进行了研究,并研究了不同饥饿程度对广聚萤叶甲成虫取食选择不同寄主植物的影响,最后对初孵幼虫取食选择进行了观察。结果表明,(ⅰ)雌虫与雄虫的平均耐饥时间和生存率之间没有明显差异；(ⅱ)除饥饿48h的处理外,其他饥饿时间处理都使叶甲的寄主选择发生显著改变；(ⅲ)广聚萤叶甲初孵幼虫对苍耳和豚草选择频次之间没有显著差异,但在识别期上有明显差异,向日葵与豚草和苍耳无论在选择频次上还是在识别期上都有明显差异,表明广聚萤叶甲初孵幼虫也具有很强的寄主专一性。7.广聚萤叶甲在我国大陆潜在分布区的预测利用CLIMEX软件对该叶甲与豚草及其近缘植物三裂叶豚草、向日葵等潜在寄主植物的适生区进行了分析和预测；采用由生态气候指数(EI)衍生的生物一气候风险指数(BCRI),分别分析了该叶甲与这三种植物同域分布的范围。广聚萤叶甲在我国大陆的潜在分布区向北可以分布到沈阳,向南可能分布到海南；而华东、华南和西南东部地区是其适宜的分布区域。广聚萤叶甲与豚草在我国北部的共同适宜边际区到达山东济南,与三裂叶豚草和向日葵的北部边际区达到辽宁沈阳,而该叶甲与这3种植物的共同适宜区主要分布在华东、华南和西南东部等地。我国北方种植的向日葵受广聚萤叶甲取食危害的风险很低。更多还原

【Abstract】 Introduction of natural enemies is an effective approach to the management of alien invasive weeds. A serious of challenges facing biological control practitioners and theoretician is how to accurately assess potential ecological risks of introducing biological control agents to non-target plants, so as to avoid new biological invasions with biocontrol. Ophraella communa LeSage (Coleoptera:Chrysomelidae), by origin of North America, was recently reported heavily attacking alien invasive ragweed, Ambrosia artemissiifolia (Compositae) in the China’s mainland. Because of not strictly monophagous, its potential risks to non-target plants are of great concern. In this paper, a series of experiments was carried out to study oviposition selection of O. communa by using both classic methods and their modifications in choice- and non-choice experiments in the laboratory and the field. The study has provided rich evidences for assessment of potential risks of O. communa to non-target plants, and useful knowledge for improvement of biological weed control practices. The following results were obtained.1. Ovipositon preferences of O. communa under indoor conditionsThe recognition time concept was used, and the time until plants were oviposited was fitted by the Cox proportional hazard model, to analyze the potential risks of oviposition as affected by feeding experiences and aging. The non-target plants for the test were A. trifida, X. sibiricum, two varieties of Helianthus annuus, and H. tuberosus, all in tribe Heliantheae of Composite. The experiments revealed a strong preference for the target weed A. artemisiifolia to the non-target test plants, but did not find significant influences of feeding experiences and aging on the preference. 2. Ovipositon preferences of O. communa under manipulated outdoor conditionsBehavioral phenomena have the potential to impact on the interpretation of test results obtained in laboratory. Thus, experiments were conducted both in a large walk-in cage in the field and in open field to observe oviposition preferences. The study showed that:(ⅰ) O. communa strongly preferred A. artemisiifolia for oviposition to non-target plants; (ⅱ) O. communa turned to A. trifida and X. sibiricum for oviposition without the presence of A. artemisiifolia, while the other non-target plants were at little risk.3. Oviposition selection of O. communa under in the open fieldA survey was made to investigate oviposition selection of O. communa for A. artemisiifolia as affected by various characteristics of host plants. Following characters were surveyed on a regular time interval and plots:plant height, plant patch size, shading, blooming, and damage levels. The factors surveyed were fitted as covariates with generalised linear mixed modeling (GLMM) to determine the major factors influencing the oviposition selection. The results showed that plant patch size, damage level and blooming have significanltly influences on oviposition selection of O. communa, while lighting and plant height, had no significanltly influences by their own. But there covariates had significanltly influences, by two-way interactions, on oviposition selection of O. communa, except the interaction between patch size and plant height.4. Oviposition strategy of O. communa for different plantsComponents of host quality (such as carbon, nitrogen and defensive metabolites) can affect potential and realized fecundity as well as oviposition strategy involving clutch size, egg size and quality in herbivorous insects. Both clutch and egg size of O. communa were greater on A. artemisiifolia than on the other non-target plants without trade-offs. However, a trade-off was detected between clutch size and egg size on X. sibiricum and H. annuus, where there was a negative linear relationship between clutch size and egg size, which suggested an oviposition strategy of either "many-small" or "a few-large".5. Oviposition strategy of O. communa for A. artemisiifoliaResource density and quality of host plant can affect oviposition strategy of herbivorous insects. Experiments were conducted to observe oviposition strategy of O. communa with different host-density patches. The results showed that the quality of host plants influenced oviposition strategy of O. communa, but the density did not. The strategy of oviposition was "many-large" at the early stage, but a trade-off between clutch size and egg size at the middle stage, and "few-small" at the later stage of oviposition. O. communa preferred A. artemisiifolia at the growing stage by laying larger egg-clutches (20.3±1.5 eggs) and egg size (0.077±0.001mm3) to the plants at seedling and flowering stages. O. communa laid a smaller clutch on the plant infected with egg-masses at lower proportion levels (0%,25%, 50% of leaves) than with the higher (75%,100%), and clutch size decreased linearly with egg-mass proportion on the plant that was exposed to oviposition. O. communa laid smaller eggs on the plant with 25% leaves infected with egg-masses than with the other egg-mass treatments, and a U-shaped curve was fitted between egg size laid and egg-mass proportion treated. The plants with lower levels of egg-mass infection (0%,25%) ran a higher hazard to be oviposited than those with higher levels O. communa laid a smaller clutch (<11.2 eggs) on the plant infected with higher levels of larvae (50%～100%) than with lower levels of larva infection (>21.6 eggs), a negative linear relationship between clutch size and larval infection proportion.O. communa laid smaller eggs (0.068±0.001 mm5) on the plant with 25% infection of larvae than on plants with the other treatments (> 0.077 mm3), and a positive linear relationship between egg size and larval infection was detected. The results suggest that O. communa performed oviposition trade-offs between egg number size on A. artemisiifolia as affected by egg- or larva-infection levels at the time of oviposition.6. Feeding preferences of O. communa adults and larvaeDeprivation of host plants can change feeding preferences of herbivorous insects. Experiments were conducted to study effects of starvationon feeding preferences of O. communa adults among different test plants. Feeding preferences of O. communa larvae among different test plants were also observed. The study showed that:(ⅰ) there were no significant differences in starvation time and survival rate between female and male; (ⅱ) except for the 48h starvation treatment, the other starvation treatments significantly affected host preferences of O. communa adults. (ⅲ) there were no significant differences in feeding choices as measured by feeding incidences (frequencies) between A. artemisiifolia and X. sibiricum in O. communa larvae, but there were significant differences as measured by the recognition time. However, there were significant differences as measured by both feeding choice frequencies and recognition time bewteen H. annuus and A. artemisiifolia, and X. sibiricum. 7. A projection of potential distribution of O. communa in China’s mainland.To assess the potential establishment of O. communa in China’s mainland, the CLIMEX software was used to predict potential habitats of the leafbeetle, its host A. artemisiifolia, and potential host plants, A. trifida and Helianthus annuus. The Bioclimatic Risk Index (BCRI), derived from the Ecoclimate Index (EI) in output of the CLIMEX analysis, was used to measure the probability of overlapping distribution between the leafbeetle and plant. The results show that O. communa has the potential to establish at locations from Shengyang, Liaoning Province in the north, through to Hainan province in the south of China, while east China, south China, and the east in southwest China, are potential suitable areas for its establishment. Suitable areas up to the north shared by O. communa and A. artemisiifolia border on Jinan, Shangdong province, while those shared by O. communa and A. trifida and H. annuus border on Shengyang, Liaoning Province. Suitable areas shared by the leafbeetle and all three plants lies in east and south China, and the eastern part in southwest China. The plantation of sunflower crop in north China is at low risk of feeding damage by O. communa.更多还原