NINE The History of Whales Read from DNA STEPHEN R. PALUMBI AND JOE ROMAN Of what value is knowledge about the history of a threatened of large consumers, including whales, turtles, and sharks and species? Is it possible to chart the future of a species without other pelagic fish, have plummeted since the advent of global information about its past? Or is knowledge about past commercial fisheries (Jackson 1997; Baum et al. 2003; Myers population sizes and carrying capacities crucial to future and Worm 2003; Roman and Palumbi 2003). As a result, our management plans? These issues are particularly relevant to current view of the oceans is missing a crucial set of organ- the management of populations of the great whales. isms. How did marine ecosystems function before the near- Populations of all the baleen whales were dramatically extirpation of large consumers? The answer depends on reduced by whaling, and unprecedented international coop- values for historical population size. eration has established a global moratorium on the hunting In this chapter we describe a new method that employs the of whales for commercial purposes until stocks recover. But analysis of genetic variation to measure the numbers of what does recovery entail? For gray whales that migrate along whales before whaling. We present published accounts of the western coastline of North America, removal from the prior methods of whale population estimation using catch United States Endangered Species list in 1994 (IUCN 1994) data and compare results from both approaches. Genetic esti- was heralded as the first time a whale population had recov- mates of whale populations are far higher than those previ- ered sufficiently so that it was no longer in danger of human- ously obtained from catch records. It is important to note that caused extinction. Removal from the list depended on the fundamental data for both methods have uncertainty; population size increases that brought gray whales to about for example, records may be missing in the whaling data, and the same levels that were thought to have existed before the mutation rate may be higher than phylogenetic analyses whaling began (Scammon 1874). For this species, the esti- suggest. This uncertainty may have a strong impact on the mated population sizes that existed before whaling have long conclusions. We end with a discussion of potential ways to served as a benchmark against which to evaluate recovery. bring divergent estimates of whaling history into accord. In addition, knowledge of the historical numbers of whales may be instrumental in determining how ocean ecosystems The Current View of the Past are constructed. Large numbers of whales may have been a key component of marine ecosystems before the oceans were For many species of whales, historical population values serve disrupted by humans. Recent reports suggest many populations as a backdrop to management. The International Whaling 102 Commission (IWC) set targets for whale population recovery through drift (Hartl and Clark 1997), where Ne is the geneti- based on the idea that a population below about 54% of its cally effective population size—roughly the number of breed- prewhaling level should be protected to enable it to quickly ing adults. This simple relationship shows that small popu- increase. For populations above this level, careful manage- lations are subject to higher levels of inbreeding and a faster ment should regulate any allowable hunt, largely based on loss of genetic diversity. the trajectory of current populations. For populations that remain at the same size for long peri- Conventional estimates of current and past numbers, ods of time, an equilibrium is reached between the addition however, seem at odds with current protection schemes. of variation generated by mutation and the loss of variation For example, it has often been quoted that fin whales deleted by genetic drift. Assuming that only mutation and (Balaenoptera physalus) in the North Atlantic Ocean had pop- drift are acting to control variation, the relationship between ulation sizes of 30,000–50,000 individuals before whaling genetic diversity (measured by the parameter ␪), mutation (Sergeant 1977). Yet current population size for this species rate per generation (measured by ␮), and effective population in the North Atlantic is estimated at about 56,000 (Bérubé size is expected to be ␪ = 4Ne␮. If the population does not stay et al. 1998). Because fin whales were extensively hunted in the same size for long periods, the equation can still apply, the nineteenth and twentieth centuries, and because they but the average genetic diversity depends on the long term were fully protected by the IWC only in 1984, it seems average effective population size (Hartl and Clark 1997). This unlikely that current populations could exceed historical relationship applies to nuclear genes. For mitochondrial values. Are fin whales fully recovered in the North Atlantic, genes, the relationship between diversity and mutation and and thus open to exploitation? Have they breached their car- population size is similar, with two modifications. The pop- rying capacity, or will populations continue to increase? Esti- ulation size is for females only, because only females trans- mates for historical population size can help answer these mit mitochondrial DNA (mtDNA) to offspring, and the factor questions. of 4 in the formula drops to 2 because mtDNA is haploid, not Similar questions can be raised for one of the best-studied diploid. As a result, for mitochondrial genes, ␪ = 2Nf␮, where whale species in the North Atlantic—the humpback whale, Nf is the effective population size for females. Megaptera novaeangliae. Extensive survey work suggests that Armed with a measure of genetic diversity for a species and there are now about 12,000 humpback whales in the North knowledge of the mutation rate, it should be possible to cal- Atlantic (Stevick et al. 2003). Using readily available published culate values for effective population size. These values will and unpublished works, including scientific papers, not reflect the size of current whale populations, but they will nineteenth-century annual reviews of the whale fishery, tend to reflect the accumulation of genetic diversity over whale charts, and a sample of whaling logbooks, Mitchell long periods of time. As a result, this diversity provides a and Reeves (1983) estimated a minimum number of hump- molecular record of past population size that is independent back whales in the western North Atlantic of about 4,700 indi- of historical whaling records. viduals. More recent work shows a record of kills totaling more The level of mtDNA diversity in a population reflects accu- than 29,000 humpbacks from the 1600s through the early mulation of mutations over about Nf generations. Thus, esti- 1900s (Smith and Reeves 2002). Although the accuracy of mates of historic numbers based on genetic diversity give a pop- these historic reconstructions has increased greatly in the past ulation size typical for the species over the past 10,000–100,000 decade, placing these data into a framework in which they can years (for species with Nf = 500–5,000 and a generation time of be related to historical numbers, and understanding the lim- 20 years), not just the past few centuries. If population size its to precision of these estimates, has been challenging. cycles over time, then the effective size can be smaller than the typical observed size. Populations that have grown steadily over time—such as humans—may have a genetic diversity that is far Genetic Approaches to Whale Populations lower than expected (Takahata 1995). By contrast, populations Recent theoretical work in population genetics has provided that have experienced a bottleneck in the past may have a a new source of data for the measurement of historical pop- higher diversity than current populations would allow. Recent ulation numbers. Patterns of genetic diversity within popu- methods in DNA analysis can sometimes allow these circum- lations are controlled by many factors, including mutation, stances to be distinguished based on branching patterns in gene selection and migration. But one of the primary determi- genealogies (see, for example, Shapiro et al. 2004). nants of genetic diversity is the long-term average population Just as for any type of historical data, historical levels of size of a species. Mutation is always the primary source of genetic diversity are subject to uncertainty. Mutation rate, genetic variation, and it tends to increase levels of DNA vari- unsampled populations, and variation between genetic loci ation at a steady, slow rate. This variation is weeded out by can affect population estimates. Other factors besides popu- natural selection against deleterious mutants or culled by lation size and mutation rate are also known to affect levels genetic drift: the process by which alleles are lost randomly of genetic diversity. The most important of these other fac- from one generation to the next, which tends to be more tors are natural selection and population structure. common when populations are small. In general, a popula- For moderate to large populations, selection generally tion loses about 1/(2Ne) of its genetic diversity per generation weeds out mutations, and so decreases levels of genetic HISTORY OF WHALES READ FROM DNA 103 variation, more quickly than drift does. So-called genetic 0.4 sweeps may reduce genetic variation in a region of the y = 0.024x0.680 genome to very low levels, while leaving levels of variation 0.35 on other parts of the genome untouched. Because the mito- chondrial genome is one long molecule, all genes in this 0.3 genome are linked, and selection on one of them can reduce 0.25 genetic variability in all of them. Selection can also increase genetic diversity when its direction varies over space or time, 0.2 if it varies with the frequency of alleles in a population, or if D-LOOP individuals with heterozygous genotypes have an advantage.