It’s one thing to be a detective. It’s another to be an artist: shifting expectations, making unlikely comparisons, causing one to consider entirely new perspectives.
Comparing elements of extant alligators and red deer to an extinct hadrosaur certainly changes how one views paleontology. There is something unifying about it, connecting traits of living species—creatures that share the world with us today—to species that died out millions of years ago. Instead of a scientific field one might put into a box labeled “the study of the past,” it becomes an increasingly complex vine weaving the past with the present. And if animals as seemingly disparate as alligators, red deer and hadrosaurs share similarities, what else among us does?
Maiasaura peeblesorum model; courtesy Dr. Holly Woodward Ballard
This connection was made all the more apparent in speaking with Dr. Holly Woodward Ballard about her background and her recent paper. Her love of dinosaurs and microscopes were a perfect match for osteohistology, a field she pursued during her Masters.
Dr. Jim Farlow and Dr. Jack Horner—both members of her PhD committee and who have experience studying the bone microstructure of alligators and Maiasaura respectively—contributed to her Maiasaura peeblesorum research. They acknowledge that comparing alligator bone growth to dinosaurs has been done before; alligator bone growth has been studied extensively.
Red deer on the Isle of Rum, however, have been studied even longer. Dr. Woodward Ballard and her colleagues found similarities to Maiasaura in their survivorship rates, as well as within their bone microstructure.
Just as the red deer in Scotland, Maiasaura seem to have experienced a high mortality rate in the first year. If, however, they survived that first year, they seemed more likely to live through sexual maturity, which may have been between 2-3 years of age. Eight or nine years marks another difficult year for both species. This is when their bodies appear to decline, or senesce, and they are at greater risk for mortality at this age. Dr. Woodward Ballard and her colleagues note that one Maiasaura tibia with 10 lines of arrested growth (“LAGs”, indicating 10 years of life) appeared to still be growing.
“We have to understand the biology of modern animals and how it works before we can make any kind of hypotheses or inferences into extinct animals,” she explained. “The most important thing I learned from this experience was that we really don’t know as much as we should know about how modern animals grow and the life history details that are stored their bone tissue.”
“It’s sort of circular in that the more we learn about modern animals to apply it to the extinct ones, the more we learn about how bone biology works, how bone grows, and that has direct applications to the medical field, to veterinary biology, and to all kinds of modern fields where bone biomechanics and that sort of thing play a big role.”
Studying bones was only part of the research. The other involved applying statistical models to the data compiled. There are advantages to so many fossils from what the authors of the paper described as a “monodominant” bonebed. As mentioned in the previous post, the Maiasaura bones originate from three bonebeds in Montana, but these bonebeds are from the same stratigraphy across 2 km. This means that the scientists can be relatively sure these animals experienced the same environmental stresses. Differences in the bones, therefore, would indicate differences within each animal instead of being caused by external factors.
And the number of tibia studied in this paper was highly significant.
“There was one paper that came out about the mortality rates–survivorship curve distribution,” said Dr. Liz Freedman Fowler of Montana State University, co-author of Dr. Holly Woodward Ballard’s paper, “and the math in that was fairly complicated. Holly wanted to make sure that she did it right, and so that’s where I came in. It is quite complicated math making sure that you get all the different steps right. Because the paper was critiquing and criticizing a previous paper that had done it wrong slightly, we wanted to use the methods of this kind of revision paper to make sure that we analyzed things appropriately.”
Dr. Liz Freedman Fowler with a painting of an entirely separate (and new!) species of hadrosaur she helped discover, Probrachylophosaurus bergei; photo by Sepp Janotta of the Montana State University News Service
“[A sample size of 50] was their suggestion,” she explained further, “because the previous histology papers that have been looking at mortality rates, they’ve been using a much smaller sample size: 10-15 individuals, [for example], which is still big for paleontology. But, you know, the smaller your sample size, the greater the chance that what you’re seeing is just random variation in your sample. Whereas when you get a larger sample size, you can be more confident that you’re more accurately representing the population.
“Normally with dinosaurs you only have maybe two or three examples of a single species. So there’s really not much you can do mathematically because there’s just not enough data to run statistics on.”
Referenced throughout their paper was one published in Paleobiology in 2011 by David Steinsaltz and Steven Hecht Orzack. The Steinsaltz/Orzack paper was a response to one published in Science in 2006.
“Based on [Steinsaltz and Orzack’s] modeling,” Dr. Woodward Ballard explained, “they recommended that the minimum sample size of 50 is what you would need for an extinct population in order to figure out what the shape of the survivorship curve is. It’s not really a hard-and-fast rule. But this is the only time that mathematicians have actually suggested a minimum number for producing statistically robust survivorship curves for dinosaurs. The fact that we were able to then meet their suggested requirements was pretty important.”
Upon first reading the paper by Dr. Woodward Ballard et al, I believed that one needed a sample of at least 50 fossils of a species in order to estimate a statistically-significant survivorship curve. But—of all numbers—why 50? And why so many when most bones of extinct species are not as abundant as those found so far for Maiasaura?
Over the course of a conversation with Dr. Steven Orzack, I learned that what he and his co-author offered was a way to decrease potential misclassification errors in statistical calculations.
In simplest terms, they were raising the bar.
The 2006 paper by Erickson et al had used a sample size of 22 different Albertosaurus skeletons to calculate a convex survivorship curve. Convex, in other words, means that the survival rates decrease with age.
Cast of Albertosaurus libratus from (appropriately for this post) Red Deer River Valley, Alberta, Canada at the Yale Peabody Museum; image taken by the author of this blog
By using computer simulation to repeatedly “resample” that estimated curve, as well as a survivorship curve that was not convex (one in which some survival rates increased with age), Steinsaltz and Orzack found that about 10% of the simulated samples of size 22 taken from the non-convex sample would look convex. Such a result would mislead a scientist to misclassify the underlying survivorship curve as being convex when, in fact, it was non-convex. When they repeated this process by more than doubling it to a sample size of 50, they discovered the misclassification error rate fell to less than 1%.
Paleontologists don’t always have access to a wealth of fossils from the same species. This is something Dr. Orzack—trained as both a paleontologist and a neontologist—knows all too well.
Images of a partial Deinonychus skeleton, discovered in Montana in 1974 by Dr. Steven Orzack and a team of Harvard researchers, now at the Harvard Museum of Natural History; images taken by the author
“I don’t have any problem with sample sizes of 22 in the sense that if that’s the best you have, that’s fine,” he said. “What would have been better is [if Erickson et al had done] the statistics better.”
“Convexity,” he stated, “is a very specific claim.”
“[There are] weaker conclusions you can make about how survival rates change with age than [those published in the paper by Erickson et al.] If you boost your sample size to 50, you have a much lower probability of saying incorrectly that there is convexity when there isn’t,” he concluded.
“Paleontology is moving in a much more mathematical and analytical direction,” Dr. Freedman Fowler explained. “ We’re trying to be more rigorous and treat it more like a modern science. That’s why we often use the term ‘paleobiology,’ instead of just ‘paleontology’ now. We’re trying to use the science and the tools of modern biology to look at how fossil organisms lived and kind of reconstruct their lives.”
And certainly, the math contained within the paper by Dr. Woodward Ballard, Dr. Liz Freedman Fowler and their colleagues is—to someone like myself—a bit overwhelming.
When speaking with Dr. Freedman Fowler, I asked her if her mathematical skills were rare within the field.
“I wouldn’t say ‘rare’,” she replied, “but it’s certainly not all of us. There are quite a lot of other paleontologists that use R and use math and things. But it’s a minority that goes in that direction.”
FIGURE 6. Survivorship curve for Maiasaura. Sample size of 50 tibiae was standardized to an initial cohort of 1000 individuals (assumes 0% neonate mortality). Survivorship is based on the number of individuals surviving to reach age x (the end of the growth hiatus marked by LAG x). Age at death for individuals over 1 year old was determined by the number of LAGs plus growth marks within the EFS, when present. Error bars represent 95% confidence interval. Mean annual mortality rates (μ^) given for age ranges 0–1 years, 2–8 years, and 9–15 years. Vertical gray bars visually separate the three mortality rate age ranges; courtesy Dr. Woodward Ballard.
“Paleontology is very collaborative because it’s such a broad and interdisciplinary field. Nobody can be an expert in everything.”
When I asked her whether the sub-fields within paleontology have always been so diverse, she responded, “It is certainly a more recent development, and that’s true for many sciences.”
“[Looking back at] papers written 50 years ago, they’re almost all single authors. They’re also much more simple. These papers were just ‘I found this new species. Here’s what it looks like.’ There wasn’t much analysis.
“But now, as all these different branches of science have grown–all the different subfields within biology and geology and chemistry–we’re getting so many more tools that we can use to analyze fossils and look at them in all these different ways. We’re also having a much larger sample size of fossils. We’re constantly out in the field collecting new specimens and that’s filling in gaps. Between two species, [for example], we now find the intermediate species. And we’re getting more complete growth series—the ontogenetic series—of animals. We’re out there finding juvenile dinosaurs and sub-adult dinosaurs and comparing them to the adult dinosaurs.
“Because we’re always adding this data, we always have more and more to work with. So we’re able to do types of analyses that we couldn’t 50 years ago. It was just impossible.”
And this paper is only the beginning. Dr. Woodward Ballard explained that she wants to “really make Maiasaura the dinosaur that we know the most about and really use it as a model to compare to other dinosaurs.”
In a moment of reflection, she said, “I get this question a lot: ‘Well, great, you’re studying dinosaurs, but what’s that going to do for me?’”
She hopes that the interest in dinosaurs will pull people into science in general, describing a scenario in which the kids—wanting to see dinosaurs—visit a museum with their parents. While there, the family may learn of other scientific discoveries, prompting even more interest in various scientific fields.
“The more we can make dinosaurs these realistic animals, [not just animals that are no longer around], I think it’s really going to get [kids] interested in science and the world around them. Being able to continue to add more information to Maiasaura, I think, is going to be the way to really draw people in.”
“The big thing for me,” she said, “is not only collecting fossils, but [also] bringing college-aged kids to Montana to see a different part of the United States, [especially those] kids who might not [otherwise] have the opportunity to be exposed to science.”
“There’s still so much that can be done with the Maiasaura bonebed,” she continued, “with Maiasaura as an animal, so [many] opportunities for outreach and scientific investigation. I spoke with Jack Horner about this during my dissertation work and afterwards; I told him that I would really like to be able to work on Maiasaura potentially for the rest of my career. He thought it was a great idea. I’ll do other research, too, but I plan to get out to Montana every summer.
“There’s just so much work that I decided to call it the ‘Maiasaura Life History Project’ and every paper that comes out will just be adding to what we already know about Maiasaura.”
At this time, there is no overall funding for the project. Dr. Woodward Ballard is currently writing grant proposals for future expeditions.
Dr. Holly Woodward Ballard; photo by Dr. Karen Chin, courtesy of Dr. Woodward Ballard
- Maiasaura, a model organism for extinct population biology: a large sample statistical assessment of growth dynamics and survivorship; Holly N. Woodward, Elizabeth A. Freedman Fowler, James O. Farlow, John R. Horner, Paleobiology, October 2015
- Statistical methods for paleodemography on fossil assemblages having small numbers of specimens: an investigation of dinosaur survivorship rates; David Steinsaltz, Steven Hecht Orzack, Paleobiology, Winter 2011
- Largest dinosaur population growth study ever shows how Maiasaura lived and died, Montana State University, MSU News Service
- MSU team finds new dinosaur species, reveals evolutionary link, Montana State University, MSU News Service
- Tyrannosaur Life Tables: An Example of Nonavian Dinosaur Population Biology; Gregory M. Erickson, Philip J. Currie, Brian D. Inouye, Alice A. Winn
**I need to stress that the methods used in this paper and the overall research by Dr. Woodward Ballard and Dr. Liz Freedman Fowler were extremely complex. Dr. Woodard Ballard, Dr. Freedman Fowler and Dr. Orzack graciously walked me through scientific and statistical elements that I had trouble understanding. If there are any errors in this post, they are my own.
Also, while comparisons between extant and extinct species may be normal to those in the field, it was not as dramatically apparent to me until this paper.
I would like to extend, again, an enormous THANK YOU to Dr. Holly Woodward Ballard. I would also like to extend that same thank you to Dr. Liz Freedman Fowler and Dr. Steven Orzack. It was a great pleasure and honor speaking with each of them–not to mention fun!–and I am profoundly grateful for their generosity!
I am very eager to learn more as the Maiasaura Life History Project continues!!