It’s just a little pre-digested; it’s still good, it’s still good.

A coprolite from the intestine of the Yukagir mammoth. The large twigs are from willows. From van Geel et al., 2008.

Blogging on Peer-Reviewed ResearchOn Tuesday, I wrote a short essay on the rightful place of science in our society. As part of it, I argued that scientific knowledge is distinct from the scientific method - the latter gives people the tools with which to acquire the former. I also briefly argued that modern science education (at least in the UK) focuses too much on the knowledge and too little on the method. It is so blindsided by checklists of facts that it fails to instil the inquisitiveness, scepticism, critical thinking and respect for evidence that good science entails. Simply inhaling pieces of information won't get the job done.

This assertion is beautifully supported by a simple new study that compared the performance of physics students in the USA and China. It was led by Lei Bao from Ohio State University who wanted to see if a student's scientific reasoning skills were affected by their degree of scientific knowledge. Does filling young heads with facts and figures lead to a matching growth in their critical faculties?

Fortunately for Bao and his team of international researchers, a ready-made natural experiment had already been set up for them, in the education systems of China and the US. Both countries have very different science curricula leading to different levels of knowledge, but neither one explicitly teaches scientific reasoning in its schools. If greater knowledge leads to sharper reasoning, students from one country should have the edge in both areas. But that wasn't the case.

In China, all students work to the same standard courses until the 12th grade (Year 13 in the UK). For physics, those run for five years, starting at the 8th grade. The curriculum is heavy on both algebra, conceptual understanding and problem-solving skills. In the US, physics education is more varied. It's covered in general science courses, but only one in three students take it up as a subject in its own right, and even then, just for two years.

It should come as no surprise that test scores reflect these differences. Using two standard physics tests - one on mechanics and one on electricity and magnetism - Bao tested the knowledge of 5,760 freshmen. Each was about to start a physics degree at one of four American and three Chinese middle-ranking universities.

In both tests, the Chinese students outperformed their American rivals. On the mechanics test, the average Chinese score of 86% easily trounced the American average of 49%. On the electromagnetism test, the Americans scored an average of 27% (barely better than chance) while the Chinese still averaged a respectable 66%. The Chinese students were also less varied in their scores, as you might expect of a nation that has a uniform curriculum. In contrast, the greater variability of American education led to a much wider range of results.

So the Chinese students clearly have the edge in terms of scientific knowledge. But in terms of scientific reasoning, students from both countries performed almost identically.  Bao also asked them to sit Lawson's Classroom Test of Scientific Reasoning - another standard test, but one designed to examine basic scientific skills. It doesn't require much knowledge, but instead probes for skills like wielding deductive and inductive logic, controlling different variables and testing hypotheses.

Faced with these questions, the average scores of the American and Chinese students were practically identical (74.7% and 74.2% respectively), as were their distributions. These striking similarities suggest that students from both countries share matching scientific reasoning skills, despite their vastly different educational systems.

To Bao, it's further evidence that modern science education "emphasizes factual recall over deep understanding of science reasoning". Even if this fact-focused system is realised in a rigorous way over many years, as it is in China, it still doesn't lead to vastly better scientific reasoning skills. And while Bao's study only looked at physics but there's reason to believe that the same would apply for other scientific disciplines. After all, both countries have similar differences in the way they teach biology and chemistry.

Now, you might argue that the fact that both sets of students still achieved respectable scores in the reasoning tests proves otherwise - surely if their education was bad at instilling reasoning skills, the average scores would be much lower? Well for a start, you might expect that the sample group - students who were already starting a university physics course - would have a natural edge in reasoning skills anyway.

But Bao also points out, "Their reasoning ability was developed through both formal education (or school education) and informal education (learning in real life)." His data show that different standards of education, leading to different levels of knowledge, don't translate into different aptitudes for reasoning.

So if it what we teach doesn't lead to better reasoning skills, perhaps it is how we teach it. That's an idea Bao subscribes to. He advocates more "inquiry-based learning", where teachers encourage students to discover things for themselves, acting more as a facilitator that guides the class rather than a vessel that pours knowledge into it. Obviously, this shouldn't be introduced to the exclusion of knuckling down and cramming some basic facts in, but the two approaches should complement each other. If this concluding chapter feels a bit rushed, it's because I freely admit that I have no experience in the educational sector - I'm more interested to hear the opinions of readers who are.

Reference: L. Bao, T. Cai, K. Koenig, K. Fang, J. Han, J. Wang, Q. Liu, L. Ding, L. Cui, Y. Luo, Y. Wang, L. Li, N. Wu (2009). PHYSICS: Learning and Scientific Reasoning Science, 323 (5914), 586-587 DOI: 10.1126/science.1167740

More on education: When learning maths, abstract symbols work better than real-world examples

Image: Graphs from Science; classroom by P.Morgan

Subscribe to the feed


If you want to know about the life and habitat of a woolly mammoth, there is scarcely a better place to look than in its dung. Found frozen in the permafrost or extracted from the intestines of well-preserved specimens, mammoth coprolites are fecal records of the plants which existed in the animal’s local environment and what foods that individual was eating just prior to death. Twigs, fruits, seeds, and other plant bits are common components of mammoth coprolites, but there are also signs that mammoths sometimes picked up the dung of their own species for a little extra digestive processing.

The head of the Yukagir mammoth. From van Geel et al., 2008.

Blogging on Peer-Reviewed ResearchThe Japanese pinecone fish searches for food with living headlights. This ­­hand-sized fish harbours colonies of light-producing bacteria in two organs on its lower jaw. The beams from these organs shine forward, and when night falls and the fish goes searching for food, its jaw-lamps light the way.

Elsewhere in the Pacific Ocean, the Hawaiian bobtail squid also uses luminous bacteria, but theirs act as a cloaking device. They produce a dim glow that matches the strength of moonlight from above, hiding the squid's silhouette from hungry fish below. It's a mutual relationship; the squid gets protection and it pays its residents with sugars and amino acids.

The glowing bacteria of these two animals may have different uses, but they are actually the same species - Vibrio fischeri, a free-swimming bacterium found in almost all of the world's oceans. V.fischeri isn't inherited; instead, it colonises the light organs of both fish and squid when they are young. Its challenge is to recognise the right partners among the myriad of species in the ocean, and not end up in the wrong body.

But its potential hosts, the bobtail squid and the pinecone fish, are incredibly different animals, separated by over 550 million years of evolution. How does one bacterium manage to form tight alliances with such disparate hosts?.

Incredibly enough, it does so with a single gene. Mark Mandel from the University of Wisconsin found that the strains of V.fischeri in the squid contain a gene called RscS that is missing or very different in those found in fish. RscS was the genetic innovation that allowed a fishy bacterium to set up shop in the body of a squid.

Mandel made the discovery by looking for differences between the genomes of two V.fischeri strains - MJ11 from the pinecone fish , and ES114 from the squid. MJ11 can't colonise squids, so Mandel reasoned that the gene (or genes) responsible for ES114's partnership would be unique to this strain and missing in MJ11's genome. RscC fitted the bill perfectly, and earlier studies had already shown that this gene plays a role in initial phases of the partnership.

In the ES114 strain, RscS switches on another gene called SypG, which in turn produces long polymers of sugar molecules called exopolysaccharides. These large molecules provide the support that V.fischeri  needs to produce large communities called biofilms and gain a foothold in their new host. These sugar polymers are the concrete that builds bacterial cities

The fishy MJ11 strain can also produce the same material but it obviously lacks the RscS gene that kick starts the whole process. But when Mandel donated copies of RscS to MJ11, he found that this strain not only produced biofilms, but successfully set up shop in young squid. They even formed colonies as readily as the squid's natural partner ES114.

By analysing even more V.fischeri strains from fish and squid throughout the Pacific, Mandel found two versions of RscS. One of them, RscSA, is found in all bacteria isolated from squid, but only one strain taken from fish. The other, RscSB, is only 85% identical to RscSA and much larger in size - it is only found in fish-based bacteria, and then only in about half of them.

It's the RscSA gene in particular that turns V.fischeri into a squid tenant, and it's telling that the only fish strain that could colonise squid was the only one with this smaller version. But it lost that ability when Mandel interfered with the gene, proving that RscSA is both necessary and sufficient for squid settlement.

By constructing a family tree of V.fischeri strains, Mandel managed to reconstruct the evolution of this bacterium, as it jumped from host to host. He showed that strains carrying either version of RscS form a tight-knit and exclusive family. RcsS almost certainly turned up in the V.fischeri genome just once, and was passed down from bacterium to bacterium. In the genomes of modern bacteria, the gene is still always found in the same place.

RscS's origins are unclear. It's unlikely to be an altered copy of another V.fischeri gene or a mash-up of several genes, for it bears no resemblance to any other stretch of DNA in the bacterium's genome. Mandel thinks it was probably a loan from another species - just one of the many genes that bacteria transfer so willingly and so regularly from one to another. 

Wherever it came from, the advent of RscC turned the bobtail squid into a potential host for V.fischeri. The RscSA version was the first to develop and strains carrying this could colonise both hosts. At some point, Mandel thinks that RscSA evolved into the larger RscSB within the bodies of fish. The gene adapted in some way to this specific host and in doing so, lost its compatibility for squid.

This story seems complicated but at its heart is a truly remarkable event. The addition of a single gene, RscS, changed the host range of a bacterium, not because the gene interacts with the host in any way, but because it unlocked abilities that V.fischeri already had.

RscS is a "regulatory gene", an executive that controls the actions of many minions - in this case, it took over command of existing biofilm-producing genes. The bacterium was already fully equipped with these genes and was almost certainly regulating them in a different way. All RscS did was to deploy them differently. How exactly this reprogramming allowed the bacteria to partner with squids is unclear, but that's a detail for a future study.

Mandel's study is one of the first to lay out the genetic forces that solidify a partnership between a bacterium and an animal. These alliances are common (think humans and our gut bacteria) but we know precious little about what makes them tick. Nor are we clear on the genetic factors that make disease-causing bacteria picky about their hosts. Why, for example, does the "Typhi" strain of Salmonella enterica affect only humans, while the Typhimurium strain (whose genes are 97% identical) infects mice and many other animals?

We probably shouldn't be surprised at that. If the story of the squid, the fish and V.fischeri tells us anything, it's that tiny genetic changes can open massive new doors for bacteria.

Reference: Mark J. Mandel, Michael S. Wollenberg, Eric V. Stabb, Karen L. Visick, Edward G. Ruby (2009). A single regulatory gene is sufficient to alter bacterial host range Nature DOI: 10.1038/nature07660

More on symbiosis:

Images: Pinecone fish by Opencage; bobtail squid by William Ormerod

Subscribe to the feed


A male woolly mammoth recently discovered in northern Yakutia, Russia ate feces shortly before it died. As reported in a 2008 issue of Quaternary Research by an interdisciplinary team of scientists, the “Yukagir mammoth” was well-preserved enough that paleontologists were able to extract a coprolite from its lower intestine and sort through the plant-rich material. Traces of willows, daisies, sorrel, sedges, rushes, sweet-grass, cinquefoils, and other plants indicative of an open, grassy landscape were found, but so was a fungus – Sporormiella – which is known to grow on deposited feces. Since the fruiting bodies of this fungus would have taken about a week to develop, and there was no sign of bile acids in the coprolite which would have signaled that the ingested dung had come from another species of animal, it appears that the Yukagir mammoth ate a one-week-old mammoth patty just prior to its death.

(In investigating the mammoth carcass, the paleontologists also found evidence that the Yukagir mammoth had back problems. The scientists proposed that it had reactive spondylarthropathy, which they said was “most probably associated with inflammatory bowel disease.”)

The woolly mammoth coprolite found in northwestern Alaska. From van Geel et al., 2010.
I don't normally post videos here, but when it's David Attenborough doing the talking, I'm more than happy to make an exception. This man is a legend, and has done so much to promote the majesty of nature to the entire world. He's now finished with big mega-series, but not with film-making. On Sunday, February 1st, the BBC airs his new one-off programme "Charles Darwin and the Tree of Life." In it, Sir David candidly speaks about his views on Charles Darwin, the vital role of evolutionary theory in linking man to the natural world, and how the attitudes espoused by the Book of Genesis have contributed to the devastation of the natural world. Have a look at an interview I did with Sir David last year, and of course, at the video below.

The Yukagir mammoth was not the only one to engage in coprophagy. Earlier this summer some of the same scientists reported another incident of feces-eating among woolly mammoths based upon their observations of  a coprolite found with a partial mammoth skeleton in northwestern Alaska. As described in Quaternary Science Reviews, this dung ball contained plants representative of the open, dry “mammoth steppe” habitat preferred by these animals, as well as the fungi Sporormiella and Podospora conica. And, like the Yukagir mammoth coprolite, no bile acids were found in the coprolite, meaning that the Alaska mammoth also ate a coprolite which had been dropped by a mammoth (itself or another) a week or so before.

The exact reasons why these mammoths ate dung is difficult to discern, but this behavior is not as unusual as it sounds. Young herbivorous mammals sometimes populate their guts with the appropriate digestive bacteria by eating feces, and herbivore droppings can be rich sources of plant material and fermentation products. Even living elephants have been observed to consume dung from time to time, but how regularly mammoths did so is unknown. Perhaps, the scientists behind these reports speculate, the mammoths studied so far were in a state of nutritional stress and turned to coprolites as an easy source of food shortly before their deaths. Then again, perhaps paleontologists have just missed the clues that mammoth coprophagy was regular behavior. Further sampling and research will be required to find out just how often woolly mammoths consumed dung, but at least these two studies have given scientists an idea of what to look out for while picking through mammoth shit.


VANGEEL, B., APTROOT, A., BAITTINGER, C., BIRKS, H., BULL, I., CROSS, H., EVERSHED, R., GRAVENDEEL, B., KOMPANJE, E., & KUPERUS, P. (2008). The ecological implications of a Yakutian mammoth’s last meal Quaternary Research, 69 (3), 361-376 DOI: 10.1016/j.yqres.2008.02.004

van Geel, B., Guthrie, R., Altmann, J., Broekens, P., Bull, I., Gill, F., Jansen, B., Nieman, A., & Gravendeel, B. (2010). Mycological evidence of coprophagy from the feces of an Alaskan Late Glacial mammoth Quaternary Science Reviews DOI: 10.1016/j.quascirev.2010.03.008

Post-script: As described by Sean, and as seen in the video below, some captive elephants go straight to the source when interested in a stinky snack:

For more on coprolites:

Fossil feces from an Indiana sinkhole preserve traces of a meat-eater’s last meal

Unique fossils record the dining habits of ancient sharks