212万年前，黄土高原已现人迹

因为不能截图，就粘贴几段话吧。

“线粒体DNA用于研究母系遗传的前提在于，线粒体DNA不容易重组和改变。但分子人类学的理论还远未成熟，分子数据应用在人类学问题时实际上常常被证实很难解释，这是分子序列的比较方法引起的后果。比如研究人种A和人种B在进行分子进化率的时候，就必须选择人种C进行参照，如果A和C之间的分子差别与B与C之间的差别类似，那么就推论A和B的分子率类似。同时由于科学依赖于未来的发现不会与现有理论产生矛盾。实际上，目前的发现也有否定线粒体遗传的假设，那么当前学说很有可能在未来20年内被完全否定。所以对待分子人类学研究的结论必须审慎，不能盲从。”

“美国科学家的新研究也显示，有关线粒体的一个关键科学假设可能存在问题，使得追溯人类母系祖先所用的“分子钟”不准确。美国哈佛医学院科学家说，mtDNA分子也会发生DNA片断交换和重组，这与此前人们所认为的不同。该成果发表在新一期美国《科学》杂志上，可能对以前的一系列科研成果造成冲击，涉及人类的进化、原始人类的迁徙，乃至各种人类语言之间的关系。”

“The molecular phylogenetics is based on quantification substitutions and then comparing sequence with other species, there are several points in the process which create errors. The first and greatest challenge is finding "anchors" that allow the research to calibrate the system. In this example, there are 10 mutations between chimp and humans, but the researcher has no known fossils that are agreeably ancestral to both but not ancestral to the next species in the tree, gorilla.”

“There are several problems not seen in the above. First, mutations occur as random events. Second, the chance that any site in the genome varies is different from the next site, a very good example is the codons for amino acids, the first two nt in a codon may mutate at 1 per billion years, but the third nt may mutate 1 per million years.”

“Statistically one can assign variance based on the problem of randomnicity, back mutations, and parallel mutations (homoplasies) in creating an error range.“

“Time to most recent common ancestor (TMRCA) combines the errors in calibration with errors in determining the age of a local branch.”