Further to our recent discussion of the decline of natural history relative to ecology (starts here), frequent commenter Joachim notes that, according to Google Ngrams, natural history’s heyday was in the late 18th and early 19th centuries. It started declining well before ecology took off, and actually has been more or less stable since the rise of ecology. This suggests that “natural history” declined because science started becoming increasingly professionalized and lab-based during the 19th century. So if the decline in prestige of natural history bothers you, you should probably be less upset with ecologists than with Herschel and Whewell. Botany peaked at around the same time as natural history. But zoology’s peak seems to have been rather later, perhaps in the early 20th century, although zoology’s popularity apparently has changed relatively little over time.
Don’t take this too seriously. All the usual caveats about (over)interpreting Ngrams apply. But I thought it was intriguing enough to be worth posting on.
Dynamic Ecology 
Maybe the what we call the studies have changed but the we are still mostly conducting (or at least publishing) observations and experiments about single species.
This from a broad survey of the ecological literature in PlosOne published this month (here):
“Ecology is mostly a study of single species. Most of the ecological
research focused on the demography, physiology and distribution
of single species. The proportion of single-species studies has
slightly decreased in the past three decades, but still consists of
more than 60% of the studies.”
And
“Observation and experiment were by far the predominant tools
of ecological study, together accounting for 80% of the research;
these proportions did not change over time. Interestingly,
modelling (,12% of all studies), is no more common today than
it was thirty years ago, despite a drastic increase in the availability
of modelling tools during this period. Data-analysis became a
more common research tool”
“Maybe the what we call the studies have changed”
Sure. Though I’m not sure that I’d call most observational and experimental studies of single species “natural history”. At a guess, I suspect most of them would be better labeled “population ecology” or “conservation biology” or “physiological ecology” or “evolutionary ecology” or whatever.
That modeling isn’t any more common today than it was 30 years ago will come as a shock to many people, I think (though personally I’m not surprised). That sounds like something worth blogging about, thanks very much for the heads up on that paper.
Looking at the resource books of some of my Ngrams (all caveats about ngrams admitted), I got the hunch that the term ecology took off in the 1960s, because it escaped its scientific enclosure and became a household term in the general public. The search for an equal take off of other terms such as “silent spring” or “environmental crisis” turned out to be a failure.Maybe Ngrams is not able to search form compound terms. But the the term pollution rockets even steeper than ecology around the same time, making the increase in the term ecology look shallow in comparison , whereas it used to look very steep in comparison to zoology and botany or natural history for that matter.
http://books.google.com/ngrams/graph?content=ecology%2Cpollution&year_start=1900&year_end=2000&corpus=15&smoothing=3&share=
Adding evolution as a term in google ngrams makes a big difference. Talking about natural history, isn’t that the sum of ecological and evolutionary research?
This catapults us back to an earlier series of topics on predictions in ecology on which I wanted to answer. My apologies for the lenghty off-topic reply.
History, the chain of consecutive events, or for that sake, natural history, cannot be foretold. It can only be explained a posteriori because we can, step by step, retrace each event in a probabilistic sense. Using probabilistic metrics, we can provide average expectations and (assumption-prone) confidence intervals, supported by what our simulations and models indicate. But that does not allow us to predict the future of individual populations or ecosystems; it only allows us to put forward hypotheses.
Even in physics, considered an exact science, we are unable to completely predict the position of more than three interacting celestial objects, even if we know their current position and speed exactly (known as the n-body problem). From this problem, chaos theory emerged, showing that even in physics long-term prediction in general is impossible. Chaos theory shows that the deterministic nature of a system does not make it necessarily predictable.
Models, however, can be excellent thought experiment, to test how sensitive x and y are to changes in P and Q, but they are not predictions. A “prediction” in ecology is most of the time done à posteriori: in a historical system, variable A “predicted” the behaviour of variable B for X percent.
It may seem like a semantic discussion to argue that we can make projections using models and simulations whereas we cannot make predictions. The purpose of predictions is to forecast the future as accurately as possible. A projection, on the other hand, wants to show that a certain change in P and Q gives rise to a change in x and y. The simulation proves a theorem about the resulting behaviour of x and y.
As shown here before, ecological studies typically explain less than five percent of the investigated variance. This does not necessarily mean that ecologists are bad scientists, it may also mean that the real world is complex and unpredictable. There’s plenty of examples of plankton mescocosm experiments in which a dozen replicates were inoculated with identical communities, and the resulting communities a few months later had diverged completely, due to drift, chance events, stochasticity … A really good and complex model may likely fit the average response over all mescocosm, but that’s not what we (or policy makers) are interested in most of the time.
We don’t want to make projections about the average response of a thousand parallel earth-like systems to climate change, we want to predict the exact response of this planet to the human CO2-pump. When we go for generalities, it’s because we can’t do better than that! We are immensely constrained by contingencies to predict the behaviour of individual systems. All we can do is replicate the simulations a thousand times, filtering out the effect of contingencies (and plain chaotic changes) in each replicate run. We can’t tell which run is the right one, because there isn’t one. Chaos theory shows that the deterministic nature of a system does not make it predictable. And physics problems such as the n-body problem occur in extremely simple systems compared to the complexity of ecological interactions among individuals that often take (from an evolutionary point of view) wrong decisions (hence being unpredictable). Darwin understood this already, when he wrote “Throw up a handful of feathers, and all must fall to the ground according to definite laws; but how simple is this problem compared to the action and reaction of the innumerable plants and animals which have determined, in the course of centuries, the proportional numbers and kinds of trees now growing on the old Indian ruins!”. So it’s not that we don’t want to make predictions about the behaviour of individual systems, we just can’t.
Although the most important “mission” of ecologists and evolutionary biologists (natural historians) may be to search for predictions in inherently unpredictable systems, it is much more important to convince policy makers that we must live with uncertainties. Hence the precautionary principle.
Pingback: Friday links: how ecological research has (and hasn’t) changed in the last 30 years, and more | Dynamic Ecology
Pingback: Stats vs. scouts, polls vs. pundits, and ecology vs. natural history | Dynamic Ecology