Department of Plant Pathology, Departments of Forestry and Statistics, and Agricultural Research Service, U.S. Department of Agriculture, University of Wisconsin-Madison, Madison, Wisconsin 53706.
Appl Environ Microbiol. 1982 Sep;44(3):695-700. doi: 10.1128/aem.44.3.695-700.1982.
Total populations of epiphytic bacteria and selected components thereof were determined on sets of 24 to 36 individual leaves (corn, rye) or leaflets (snap bean, soybean, tomato) of field-grown plants by washing and dilution plating. In general, levels of component populations (e.g., bacteria that are ice nucleation active) were quantitatively more variable from leaf to leaf within a set than were total epiphytic bacterial populations. Populations of a given component frequently varied by a factor of 100 to 1,000 within a set of leaves. Total bacterial populations usually varied by a factor of about 10. For each set of leaves, total and component epiphytic bacterial populations were found to approximate a lognormal distribution by the Shapiro-Wilk test for normality. Due to the lognormal distribution of epiphytic bacterial populations, estimates of population size based on the common practice of using bulked samples (wherein several leaves are washed together) will overestimate the population median by a factor of approximately 1.15sigma. From the known probability distribution of bacterial populations, the frequency with which a given bacterial population size is met or exceeded on individual leaves can be estimated. If the bacterial component is phytopathogenic, the frequency estimates could be used to relate quantitatively pathogen populations and disease incidence.
通过洗涤和稀释平板法，在大田种植的植物的 24 到 36 个单个叶片（玉米、黑麦）或小叶（豇豆、大豆、番茄）上，确定了附着细菌的总种群及其选定成分的种群。一般来说，成分种群（例如，具有冰核活性的细菌）的水平在一组内从一个叶片到另一个叶片的变化是定量的，而总附着细菌种群则更为多变。在一组叶片中，给定成分的种群经常变化 100 到 1000 倍。通过 Shapiro-Wilk 正态性检验，发现每组叶片的总附着和成分附着细菌种群都近似于对数正态分布。由于附着细菌种群的对数正态分布，基于常见的使用混合样本（即一起洗涤几个叶片）的方法估算种群大小，会使种群中位数的估计值高估约 1.15sigma。根据细菌种群的已知概率分布，可以估计在单个叶片上遇到或超过特定细菌种群大小的频率。如果细菌成分是植物病原体，那么频率估计可以用于定量地关联病原体种群和疾病发病率。
