Hoelzel, A.R. Impact of population bottlenecks on gentic variation and the importance of life-history; a case study of the northern elephant seal. Biological Journal of the Linnean Society. 68:23-39

Background: The northern elephant seal (Mirounga angustirostrus) went through a major bottleneck in the late 19th century. In 1892 it was believed that seven of the last eight remaining individuals of the species were killed. By 1922 the population had grown again to 350 seals and up to 120,000 seals in 1980.

Hoelzel investigates genetic diversity in the post bottleneck population of northern elephant seals (NES). The author compares post-bottleneck measures of variation to pre-bottleneck, estimated by current levels of genetic diversity in southern elephant seals (SES), a closely related species which did not undergo a bottleneck like the northern population. The NES and SES genetic diversity was compared for mtDNA, allozymes, minisatellites, microsatellites and MHC. In most cases the NES had much less genetic diversity. NES had no allozyme variation (NES H=0, P=0, SES H=.034, P=.114). Two haplotypes were found in NES in contrast to 24 for SES.
In another paper author and others create a model recapitulating population growth after a bottleneck. The model is briefly reviewed in this paper. Model uses life-history data, specifically, age-specific mortality and reproduction and assumes the population exhibits density independent growth. The simulation results showed the size of bottleneck was important (the less individuals the greater the probability of extinction) and that duration of bottleneck was also an important consideration (the probability of extinction was greater for bottlenecks of 10 years than those of 1 year duration). Life-history data for NES used for these simulations was obtained from a long-term study conducted by LeBoeuf and others (1974, 1988, 1994).
The simulation model was used to test predictions about NES. Author (again in previous paper) used the simulation to back-calculate severity of the bottleneck using genetic (haplotype data) and demographic data. The simulation was reiterated 500 times. Hoelzel explains how haplotypes correspond to number of matrilines. Model determines differential survival and reproduction of martrilines. Given year of bottleneck (1884) and population size in 1960 (15000 seals), the results showed that the most likely event was a bottleneck of less than 20 seals, consistent with historical beliefs.
Authors then discuss impact of allelic diversity and heterozygosity at diploid nuclear genetic loci (again from previous paper). Heterozygosity is estimated by Ht=(1-1/2Ne)H t-1 where Ht = heterozygosity in current generation and H t-1= heterozygosity in previous generation. Authos applied this to lions in Ngorongoro crater (fell to bottleneck of 10 individuals). Serengetti lions (3000 lions) were used as "pre-bottleneck" estimates. The estimated H , 0.023, was very close to the observed value, 0.022. Hoelzel in this paper then uses the same method for nuclear loci in NES. Again, simulation results were in close agreement with observed levels.

Some of aouthors concluding points--
The duration of bottleneck related to growth rate of a population and effective size of population is related to reproductive skew. Reduced reproductive rate increases length of bottleneck and thus final level of heterozygosity reduced. Also consider male and female reproductive success. NES is highly polygynous. Hoelzel shows that the impact on heterozygosity would be much less if NES were monogamous.
Simulations consider only neutral models. Selection could decrease heterozygosity even further, or conversely, maintain heterozygosity.