Hi, Arora. I'm not entirely certain I'm reading you correctly, so this may be an unnecessary comment...but it might help some others who want a bit of clarification. There are three common types of DNA tests we see for genealogical/ancestral purposes. These look at mtDNA, yDNA, and autosomal DNA.
Mitochondrial DNA isn't human nuclear DNA at all; the mitochondria are organelles that exist in the cells of all animals...an average of about 100 in every human cell (well, except red blood cells; different story). Everybody has 'em; we've gotta have them to live. But as you say, the inheritance path is only from the mother: it follows the far right-hand side of the ahnentafel chart. Every child will have his or her mother's mtDNA, but only the daughters can pass it down to future generations. No recombination or mixing occurs with mtDNA, so there is never any kind of dilution. The mtDNA molecule is positively tiny and there's very little room for mutation. That's why mtDNA typically can look farther back in time than any other form of DNA, easily tens of thousands of years. It isn't very useful in matching a recent generation to a recent generation--again, because there's not much room for mutations to happen, not much variation--but your matrilineal ancestor from 10,000 years ago would have essentially the same mtDNA that you do today.
On the left-hand side of the ahnentafel chart is yDNA. The Y-chromosome is passed from father to son to son. The Y-chromosome escapes crossover during meiosis when the DNA replicates, and it never combines with the mother's DNA. Like mtDNA, it can look thousands of years back in time because the only changes that occur to it do so via mutation, not any type of dilution. The Y-chromosome has over 3,500 times as many base pairs of DNA as mtDNA, so there's much more room for mutation to occur without affecting the viability of the organism (though even that much larger than the mtDNA molecule, the Y-chromosome represents less than 2% of all the DNA in a genome). That makes yDNA unique in genealogy because it not only goes back many thousands of years with only minor changes along the way, but it's also useful in genealogical matching of recent generations.
Autosomal DNA does dilute; for autosomal DNA I'll mean the 22 pairs of autosomes and not include the X-chromosome, because the X behaves in a hybridized fashion and can be difficult to interpret/predict in some scenarios. And atDNA doesn't just dilute: before it mixes with the other parent's DNA it undergoes what's called crossover during meiosis, as the gamete is formed. Crossover sort of scrambles each parent's DNA so that the 50% that goes into the gamete isn't going to be exactly the same DNA that goes into the next gamete. Voila! Genetic diversity, and how species are able to survive even in situations with limited population pools (think isolated Pacific islands with endogamous populations, for example). The whole design mixes so well that even 4th cousins--sharing the very same 3g-grandparents--have less than a 50/50 chance of showing any measurable shared autosomal DNA at all.
Mitochondrial DNA and the Y-chromosome are very accurate indicators of early and ancient origins...but only for those particular and specific matrilineal and patrilineal lines. I'm in a project right now that is closing in on being able to pinpoint our shared male line--which includes, currently, five different surnames--to being in Ireland or southwest Scotland before surnames were adopted, in the timeframe of around 100 BC to 1000 AD (the line likely came from the Iberian Peninsula before that, during the Atlantic Bronze Age, and with several more Big Y-500 tests pending we hope to narrow the time farther onto the AD side of the spectrum; we think we can get it to within 500 years). Again, though, mtDNA and yDNA each represent a very small fraction of the DNA that makes up our genomes, and a tiny fraction of our ancestral grandparental genetic contributions.
If someone is basing biological Native American ancestry on mtDNA or yDNA, the DNA tests will be accurate and the results are what they are. Nothing changed enough genetically along those matrilineal and patrilineal lines over the course of thousands of years to make a difference. Autosomal DNA is a whole 'nuther matter entirely. And that's where the vast majority of genealogists are looking today for "ethnicity." There are a whole lot of assumptions, genotype modeling algorithms, fuzzy math and, well, smoke and mirrors going on there given the current state of the technology. You can read some of my recent, curmudgeonly opinion about it here.
I love that the "trade your lederhosen for a kilt" advertising trope has led to millions getting their DNA tested when they otherwise may not have. I abhor that the term "ethnicity" continues to be used; that the results are touted to be accurate; and that people then assume the results to be accurate...either that or pronounce all DNA testing as a sham because they disagree with the results.
The "ethnicity" results from that $99 autosomal DNA test should come with a warning label: For Entertainment Purposes Only. So kinda back to your point, autosomal "ethnicity" results have no use as evidence.
