Collapse: How Societies Choose to Fail or Succeed Read Online Free PDF Page B

populations at another site as unmanipulated controls. I had to use different methods. Similar methodological problems arise in many other areas of population biology, as well as in astronomy, epidemiology, geology, and paleontology.
    A frequent solution is to apply what is termed the "comparative method" or the "natural experiment" —i.e., to compare natural situations differing with respect to the variable of interest. For instance, when I as an ornithologist am interested in effects of New Guinea's Cinnamon-browed Melidectes Honeyeater on populations of other honeyeater species, I com pare bird communities on mountains that are fairly similar except that some do and others don't happen to support populations of Cinnamon- browed Melidectes Honeyeaters. Similarly, my books The Third Chim panzee: The Evolution and Future of the Human Animal and Why Is Sex Fun?
    The Evolution of Human Sexuality compared different animal species, espe cially different species of primates, in an effort to figure out why women (unlike females of most other animal species) undergo menopause and lack obvious signs of ovulation, why men have a relatively large penis (by animal standards), and why humans usually have sex in private (rather than in the open, as almost all other animal species do). There is a large scientific litera ture on the obvious pitfalls of that comparative method, and on how best to overcome those pitfalls. Especially in historical sciences (like evolutionary biology and historical geology), where it's impossible to manipulate the past experimentally, one has no choice except to renounce laboratory experi ments in favor of natural ones.
    This book employs the comparative method to understand societal collapses to which environmental problems contribute. My previous book (Guns, Germs, and Steel: The Fates of Human Societies) had applied the comparative method to the opposite problem: the differing rates of buildup of human societies on different continents over the last 13,000 years. In the present book focusing instead on collapses rather than on buildups, I compare many past and present societies that differed with respect to environmental fragility, relations with neighbors, political institutions, and other "input" variables postulated to influence a society's stability. The "output" variables that I examine are collapse or survival, and form of the collapse if a collapse does occur. By relating output variables to input variables, I aim to tease out the influence of possible input variables on collapses.
    A rigorous, comprehensive, and quantitative application of this method was possible for the problem of deforestation-induced collapses on Pacific islands. Prehistoric Pacific peoples deforested their islands to varying de grees, ranging from only slight to complete deforestation, and with societal outcomes ranging from long-term persistence to complete collapses that left everybody dead. For 81 Pacific islands my colleague Barry Rolett and I graded the extent of deforestation on a numerical scale, and we also graded values of nine input variables (such as rainfall, isolation, and restoration of soil fertility) postulated to influence deforestation. By a statistical analysis we were able to calculate the relative strengths with which each input vari able predisposed the outcome to deforestation. Another comparative ex periment was possible in the North Atlantic, where medieval Vikings from Norway colonized six islands or land masses differing in suitability for agri culture, ease of trade contact with Norway, and other input variables, and also differing in outcome (from quick abandonment, to everybody dead af-
    ter 500 years, to still thriving after 1,200 years). Still other comparisons are possible between societies from different parts of the world.
    All of these comparisons rest on detailed information about individual societies, patiently accumulated by archaeologists, historians, and other scholars. At the end of this