The laws for the inheritance of animal mitochondrial DNA differ from those revealed for nuclear DNA. In contrast to nuclear genes, animal mitochondrial DNA (mtDNA) is predominantly inherited through the maternal line and is typically assumed to be nonrecombining. The absence of both paternal transmission (hereafter: paternal leakage) and heterologous recombination of mtDNA are assumed to be key characteristics of mitochondrial DNA inheritance, which has enabled evolutionary models to be much simpler than those needed for the interpretation of nuclear DNA. However, recent revelations of paternal leakage in the animal kingdom challenge our current knowledge about mtDNA inheritance and the utility of mtDNA as a molecular marker. The occurrence of paternal leakage potentially introduces new haplotypes into populations and therefore impacts on the interpretation of mtDNA analysis. To date, it is unclear whether the documented cases of paternal leakage are exceptions to the general rule or if these events occur more frequently than so far believed. If this event occurred at a measurable frequency, it is vital to implement such data into models of mtDNA evolution to improve the accuracy at which evolutionary relationships and times of divergence are estimated. In this thesis, I aimed to provide an insight into the broader patterns of mtDNA inheritance using chinook salmon as a model organism. I first sought to delimit the frequency of paternal leakage in chinook salmon and further investigated two major mechanisms which are believed to limit paternal leakage: The many-fold dilution of paternal mtDNA by maternal mtDNA upon fertilization and the genetic bottleneck mtDNA is believed to be exposed to during early developmental stages. A screen of roughly 10.000 offspring did not reveal the presence of paternal mtDNA within these samples delimiting the maximum frequency of paternal leakage in this system to 0.18% (power of 0.95) and 0.27% (power of 0.99), suggesting that the occurrence of paternal leakage is most likely an exception to the general rule. To infer the dilution of paternal mtDNA upon fertilization, I employed real-time PCR and determined the mtDNA content of salmon spermatozoa and oocytes to be 5.73 ± 2.28 and 3.15x109 ± 9.98x108 molecules per gamete, respectively. Accordingly, the estimated ratio of paternal to maternal mtDNA in zygotes is 1:7.35x108 ± 4.67x108. This estimate is 3 to 5 orders of magnitude smaller than the ratio revealed for mammals. Consequently, and if the dilution acts as an efficient barrier against the transmission of paternal mtDNA, paternal inheritance of mtDNA per offspring will be much less likely in this system than in mammals. To estimate at what probability the diminutive contribution of paternal mtDNA in zygotes is potentially inherited to offspring, I determined the size of the bottleneck acting on mtDNA during both embryogenesis and oogensis by examining the transmission of mtDNA variants to offspring and oocytes within a pedigree of heteroplasmic individuals. The number of segregating units (mtDNAs) between a mother’s somatic tissue and oocytes was estimated to be 109.3 (median = 109.3; 62.4 < NeOog < 189.6; 95% confidence interval) and from a mother’s soma to offspring’s soma 105.4 (median = 105.4; 70.3 < NeEmb < 153.1; 95% confidence interval). Detected variances in allele frequency among oocytes were not significantly different from those in offspring, strongly suggesting that segregation of mtDNA occurs during oogenesis with its completion before oocyte maturation. However, considering a ratio of roughly 1:7.35x108 for paternal to maternal mtDNA in zygotes and that approximately 109.3 (NeOog) of the mitochondrial genomes present in zygotes are ultimately inherited to offspring, the probability for paternal mtDNA to be transmitted to offspring is in round terms 1.0x10-11/paternal mtDNA molecule. In summary, the results presented in this thesis document the presence of efficient barriers to prohibit the inheritance of minor allele contributions, such as paternal mtDNA, to offspring. These results strongly suggest that paternal leakage is an exception to the general rule. Furthermore, in comparison to studies undertaken in mammals, my results indicate that mechanisms in place to prevent paternal leakage may be unequally efficient among different animal taxa, reflecting differences in life traits, such as gamete morphology, gamete investment and reproductive strategies. Nonetheless, by the means of the dilution effect in zygotes and the genetic bottleneck during oogenesis, the occurrence of paternal leakage might be simply a quantitative phenomenon and cannot be excluded per se. The increasing number of documented cases of paternal leakage clarifies that its occurrence must be considered when applying mtDNA as a genetic marker. Furthermore, for species in which mtDNA inheritance can be confirmed to be purely random, theoretical frequencies of paternal leakage can be inferred and potentially implemented into models of mtDNA evolution.