在茶树田间试验中,经常要进行大量的观察鉴定工作。对观察所得的数据只有用生物统计的方法,才能够正确地显示出研究对象各种状况之间的严格关系,并由此得到明确简洁的结论。在实践中,我们不可能对处理的所有材料(统计上谓之全及总体)都进行观察,必须采用取样的方法,即选取一部份个体来概括总体,从而达到以样本了解大群的目的。然而,不管什么样的田间设计和取样方法,由于样本与总体总存在着差异,因此几乎所有试验都会出现取样误差。而取样误差的大小除了决定于总体本身的变异程度和取样的多少外,又往往与样本中的极值 In the field trials of tea trees, often a lot of observation and identification work. Only the biometric method can be used to observe the data, so as to correctly display the strict relationship between the various conditions of the research subjects and thus obtain the concise and concise conclusions. In practice, it is impossible for us to observe all the materials processed (statistics and population statistics). We must adopt a sampling method, that is, select a group of individuals to summarize the population so as to achieve the purpose of knowing a large group by samples. However, no matter what kind of field design and sampling method, due to the total sample and the total there are differences, so almost all the trials will have sampling error. In addition to the size of the sampling error depends on the overall degree of variation and the number of samples, but often with the sample in the extreme