Mark Twain popularized the old adage, “there are lies, damned lies, and statistics.” Statistics can be used to show just about anything. David Stockman headed Ronald Reagan’s Office of Management and Budget in the early 1980’s. He dazzled the media with statistics he used to show the virtues of Reaganonomics. Today, Mr. Stockman admits it was all a conjob. I’m not claiming the recent study entitled “Northeastern North American Pleistocene Megafauna Chronologically Overlapped Minimally with Paleoindians,” an article published in the February Volume of Quaternary Science Reviews, uses statistics in as dishonest a manner, but I do believe the results and conclusions should be viewed with suspicion.
Matthew Boulanger and R. Lee Lyman, anthropologists from the University of Missouri, compiled a list of radio-carbon dated Paleoindian sites from New York, Pennsylvania, New England, and parts of eastern Canada. They also compiled a list of radio-carbon dated late Pleistocene megafaunal remains from this same region. They used a statistical method of analysis known as Summed Probability Distribution to compare the overlap in time between the 2 lists of sites. They determined Paleoindians overlapped only minimally with megafauna in this region. They concluded that megafauna populations were already in decline before man colonized this area of the continent, and therefore suggest man may have played a role in hunting the last surviving remnants of the megafaunal populations here, but environmental change was the greater factor in their demise. I became interested in this study because I favor a protracted overkill scenario as the only explanation that makes sense for the North American megafaunal extinctions, and this paper supports environmental change instead as a greater factor in the extirpation of various megafauna in this particular region. I’ve read and reread this paper and believe there are several logical flaws in the interpetation of the results of this study.
The authors of this study admit “the youngest dated extinct animal and the oldest dated (anthropogenic) artifact do not represent the last living individual (animal) or the first (human) colonizer.” This means the theoretical statistical overlap between man and megafauna understates the real world overlap. Shawnee-Minisk is the oldest dated Paleoindian site in this region at 12,800 calender years BP, while the youngest mastodon specimen is dated to 12,700 BP, supporting this paper’s claim that there was minimal overlap. However, it seems reasonable to assume that humans not yet detected in the archaeological record colonized this region at least 150 years before Shawnee-Minisk, and it also seems reasonable to assume the actual last mastodon died at least 150 years after the youngest dated specimen known from the paleontological record. Based on these reasonable assumptions, the actual time of overlap between humans and megafauna would be 400 years–plenty of time for humans to have had a major impact on megafauna populations. At the very least, these reasonable assumptions invalidate the title of this paper. Actually, I think adding 150 years to both sides of this spectrum is a conservative assumption and the real overlap is probably much longer.
This is a surprisingly not too sloppy facsimile of a graph from the paper discussed here. It neatly sums up the data. The graph is known as a Summed Probability Distribution Curve. Click to enlarge. Blame my poor artistic ability, if it’s not an 100% exact copy.
Figure 4 in the above image illustrates the results of the study. The graph shows a perfect correlation between a sudden decline in megafauna specimens with the rise of human archeological sites. The authors of thes study acknowledge the data could be used to support an overkill scenario. Instead, they have a bizarre interpetation of these results. They reject the possibility of overkill as the primary cause of the decline because the spike in megafauna specimens just prior to human entrance represents an increase in megafauna death rates from environmental stress, and there’s no evidence of humans exploiting megafauna in the region, so they must not have been an important factor in their extirpation here. Both of these objections are highly illogical, especially the first one as I shall explain below.
The paleontological record can’t be used to establish the rates of deaths within a population, unless a catastrophe such as a flood or volcanic eruption is involved, and this is not the case in this region. There is no way of determining there was an increase in deaths within a population because every member of that population eventually did die. The spike in megafaunal remains found that date to between 13,800 BP and 12,700 BP should be interpeted as an increase in the total megafaunal populations during that time period. Or a factor, such as higher water levels making it more likely the animals became covered in sediment, increased the odds of preservation. That it was interpeted as being an increase in the death rates due to environmental stress is illogical to the point of being ridiculous, and it’s amazing that none of the dozen or so editors who reviewed this article didn’t point out this logical fallacy.
The author’s second objection is also illogical but widely shared by archaeologists who study this era. For decades many archaeologists have rejected overkill as a reason for megafaunal extinction because they don’t see enough evidence of human-killed and/or butchered megafaunal remains. This is an unreasonable expectation. Not a single skeletal specimen of a paleoindian has ever been discovered in northeastern North America. Why do archaeologists expect to find remains of the animals they killed? In fact, only 140 mastodon specimens have ever been found in the state of New York. The mastodon specimens found in New York date from between 32,000 BP-30,000 BP and from about 14,000 BP-12,000 BP. In between these 2 time periods most of New York was covered by glacial ice and was not suitable for mastodon inhabitation. To show just how low the odds are that an archaeologist would find evidence of a human-killed mastodon, let’s estimate the total population of mastodons that lived in what’s now the state of New York in the 4,000 year time frame mentioned above. New York is 49,000 square miles but only about 1/3rd of that was suitable for mastodons which preferred wetlands. The population of mastodons was not spread evenly throughout the state, and they lived in herds, but we can safely assume an average of about 1 mastodon per 50 square miles. Based on an estimate of a total suitable habitat of 16,000 square miles this equals an average population of 320 mastodons living in the area of New York at any given time. The average lifespan of a mastodon was probably about 40 years. In a 4000 year period this equals 100 generations X 320 = 32,000 total mastodons. This means the 140 mastodon specimens found in New York represents less than 1% of the total population that ever lived during this 4000 year period. But the odds of finding human modified mastodon remains are even less than that because most of the mastodons that ever lived did not die at the hands of man.
The author’s claim that megafauna were under environmental stress between 15,000 BP-12,900 BP is not well supported by the scientific literature. They cite a study of sporomiella abundance (Robinson 2005) as evidence that megafauna populations declined well before the archaeological evidence demonstrates humans were in the region then. (Sporormiella are dung fungus spores used as a proxy to estimate megafauna populations.) However, the authors of sporomiella studies suggest their evidence supports protracted overkill as a cause of extinction because there’s little evidence of a change in plant communities at the time of megafaunal population collapse. Matthew Boulanger insists humans can’t be responsible for this decline because there’s no archaeological evidence they were in the region then. But the absence of evidence is not evidence of absence. I think Boulanger’s and Lyman’s study presents data that supports an intriguing possibility–a transient human population was attracted to the American northeast by the abundant herds of megafauna that occurred here early during the Boling-Olerod interstadial (15,000 BP-14,000 BP), but these early colonists left the region without leaving much evidence they were here after they reduced large mammal populations. Megafaunal populations rebounded between 13,800 BP-12,700 BP (as shown in Figure 4) enough to once again attract humans to the region, and this time they stayed long enough to leave archaeological evidence.
Another study cited by Boullard’s and Lyman’s paper to support their opinion that megafauna were suffering nutritional stress during the Boling-Olerod is (Yansa and Adams 2012), a very biased review of scientific literature of mammoths and mastodons that lived in the Great Lakes region during the late Pleistocene. (Note: I’m not claiming Yansa and Adams are wrong because they are biased. I freely admit to being biased in favor of protracted overkill.) Yansa and Adams believe environmental change caused the extinction of mammoths and mastodons in the midwest, but the only studies they cite that actually analyzed stress in proboscidean populations in this region were conducted by Dan Fisher, and he favors overkill over climate change. The Boling-Olerud was a warm wet interstadial that allowed a greater variety of plant foods to grow in the northeast. Megafaunal populations should have been on the increase, bolstered by an improvement in nutrition. Nutritional stress is the opposite from what one would expect to see, and there is little evidence for it. A decrease in body size and earlier maturation can be attributed to stress from being overhunted.
Climate graph of the past 150,000 years. The Pleistocene megafauna survived many drastic changes in climate including spikes of global warming. There are very few known large mammal extinctions that occurred during the transition from the Illinois Ice age to the Eemian (or Sangamonian) or during the many interstadials that occurred during the Wisconsinian Ice Age. It doesn’t seem likely that the global warming of the Boling Olerud which is so brief it doesn’t even show up on this graph could have caused the extinctions.
I also think Boullard and Lyman error by lumping all of the megafauna together. The youngest dated megafaunal specimen in their data is from a giant beaver (Casteroides ohioensis) that is at least 800 years younger than Shawnee-Minisk. That is more than a minimal overlap for this species and invalidates the title of their paper. They dismiss this specimen because the majority of megafaunal specimens are older than most of the paleoindian sites. Even if what they claim is true for mastodon, mammoth, stag-moose, and flat-headed peccary, they know it’s not true for giant beaver and they should have considered an alternate title for their paper.
The various species of megafauna had different environmental preferences. Mastodons and stag-moose should have been on the increase between 15,000 BP-12,900 BP because wetland habitat expanded, and these 2 species preferred mesic environments. Flat-headed peccaries and helmeted musk-oxen were well adapted to the arid conditions that occurred during the Younger Dryas (12,900 BP-11,000 BP). Yet both xeric-loving and mesic-loving species became extinct. There is no known environmental explanation for the extinction of animals that had such widely disparate habitat preferences. The entrance of man into the region is the only explanation that makes sense.
Boulanger, Matthew; and R. Lee Lyman
“Northeastern North American Megafauna Chronologically Overlapped Minimally with Paleoindians”
Quaternary Science Reviews February 2014