Lenski experiment and falsifiability

I’ve been occasionally following Richard Lenski’s long-term evolution experiment, but the methodology raises a certain question for me. Evolution is not observable in nature because of the long time frames involved. But like Lenski’s work, we can witness the evolution of e coli much faster.

Over some decades (let’s say 50 years), in the right laboratory conditions, we could conceivably observe well over a million generations of E. coli and billions of organisms.

Now, if we were able to witness a million generations of mammals, we’d conceivably witness some amazing feats of what evolution could achieve. If I understand correctly, echolocation in marine mammals developed over the span of a well under a million generations (~4 million years or so).

So, if the evolutionary process can achieve such marvels in the span of less than a million generations in mammals, with a far smaller population and thus fewer mutations to work with, why, exactly, do we not see unguided evolution achieving any remotely similar accomplishments over the same generational span in those organisms where we can observe them for over a million generations?

And a related question… how many generations of E. coli would we have to witness, with no significant increase in complexity (no new organelle or novel complex function) before we could legitimately say that the Darwinian evolutionary theory has been falsified? In other words, is the theory testable and falsifiable? and if so, over how many generations would we have to witness the organism remain essentially unchanged before we could say the theory has been falsified?

Bacteria have been around for 3.5 billion years. Do you think the fact that bacteria still exist falsifies evolution?


From my admittedly limited understanding, Lenski simplified his experiment such that large evolutionary changes would not be expected. Some things that lead to natural selection, such as changes in the environment, we’re absent…


Any mammal that can hear can echolocate. You can do it - try it out in a pitch-black room. You won’t be able to echolocate well, but neither could the first stem-group whales.

So, evolution didn’t produce something entirely new here - it honed a previously-existing function.


This is correct. The environment for the Lenski experiment is incredibly stable. Also, sexual reproduction is absent. Thus no directional selection, and variants have to accumulate within a lineage without swapping of alleles through sex. So it’s not easily mapped to populations under changing selection or with sexual recombination.


To echo Christy, we have to realize the end result of billions of years of evolution is: E. coli. Now found on the touch screens of McDonalds everywhere.


Fun fact: Humans generate about a mole of E. coli roughly every year (within an order of magnitude). That, despite comprising only about 0.5% of the bacteria in our gut.


Sir, I respectfully beg to differ on a few points.

Firstly, I about busted my nose trying your experiment. Perhaps I may some day learn to hear my voices echoes well enough to detect a sheer wall. However, calling this “echolocation” to me stretches the definition of the term. I suppose I could refer to rhinoceri and orangutans as “aquatic animals.”

Secondly, though, if I were to learn to “echolocate,” I would be emitting sounds from my vocal system and receiving them (and determining direction) through my ears. I am not emitting directionally focused ultra-high-frequency sounds through specially purposed organs entirely outside of my vocal system, and certainly not receiving those sounds through my teeth/jawbone. Thus to say that this is simply “honing” what already exists seems exaggerated at best. This is a complete redesign.

A ship unequipped with a fathometer could drop dynamite over the side, stick a hydrophone in the water, listen for the time difference between the explosion and the echo, and work out on paper the depth of the water based on the time difference. It can determine the water depth, albeit not “well.” But saying that all ships thus equipped have a “working fathometer” and that ships with real fathometers don’t have anything new, simply that we honed a pre-existing functionality, seems a bit of a stretch.

So, back to my basic question - how long, roughly, would we theoretically have to watch a population remain essentially unchanged, and/or how many populations would we have to watch, etc., before we could start to doubt and/or falsify macroevolutionary theory?

What I’m seeing in the replies thus far seems to suggest that it simlply is not falsifiable or empirically testable, regardless of how long we watched? No matter what experiments we did with populations, no matter how much empirical observation we gave, no matter how long (10,000,000 years from now even?) we observed animal populations never evolving new complex features, none of this would ever be sufficient to falsify the basic Darwinian macroevolutionary theory?

Fair enough… but we can still study e coli at large in nature, and still recognize it as essentially having no macroevolutionary change throughout the world across time. As such, my question still remains (or is exacerbated). We can and have essentially witnessed well over a million generations of e Coli since we began studying it, and in billions upon billions of organisms, over a million+ generations, no significant macroevolutionary change. No new organelle, nothing that seems to be the nascent form of a new organelle, etc.

But, I’m asked to believe that this process, which shows no empirical evidence of macroevolution in the countless species of bacteria we have been able to observe to some extent or another… this process that cannot seem to achieve macroevolutionary changes in billions upon billions of organisms over a million generations can achieve such marvelous feats and stunning transformations over similar generational periods with far, far fewer organisms to work with.

My basic question remains - just how many generations would need to go by, in how many and what particular organisms, wherein they remain essentially unchanged and show no macroevolutionary changes, before we should doubt the sufficiency of natural selection and variation alone to achieve these changes?

In other words, is evolution in this sense empirically falsifiable, and if so, roughly what are the constraints and particulars that would be needed in order to do so?

Humans can learn to echolocate. Check out this BBC video.

1 Like

You cannot falsify evolution by observing a population remain unchanged.

Here is an article about how you falsify evolution: https://www.newscientist.com/article/dn13675-evolution-myths-evolution-cannot-be-disproved/

I know that it’s easy to look at the bacteria and think: “Well, they don’t look like they’ve changed much over time”. But for someone with a biochemistry and microbiology background, it’s hard to suggest that they aren’t constantly evolving. Biochemically and in adaptation to their local environments, they’re churning through a lot differences. For example, Escherichia coli and Salmonella typhimurium represent two branches of enteric bacteria that diverged perhaps 100 - 160 million years ago. It has been estimated that the “E. coli and Salmonella enterica lineages have each gained and lost more than 3 megabases of novel DNA since their divergence.”(1) Compare that to an average genome size of roughly 5 megabases for these bacteria. Humans and chimps share about 98%+ of their sequences whereas variants of E. coli can differ by 1 or more megabases or roughly 20%.

(1): J Mol Evol. 1997 Apr;44(4):383-97. Lawrence JG, & Ochman H


Well, I don’t disagree in principle with Dr. Venema’s observation, and I don’t doubt I could theoretically learn how if I devoted time to it, or that those with more developed or attuned sense of hearing (as those who are born blind) could do so. In theory, a blind dog could conceivably be taught how to hear its own echo in order to avoid obstacles. But I just generally object to words or descriptions becoming so broad that they no longer keep function per their intent.

Wikipedia article, for instance, says…

If we adopt the much more broad definition of echolocation, we’ll need to say this description is erroneous and say instead that echolocating animals include all mammals that can hear. I just find that stretching the term beyond usefulness.

Well, I did the comparison as suggested, and let me ask just to clarify, as I’m not understanding the result:

If humans and chimps share 98% of their DNA, this equates, to having around 60 megabases of difference between the two species; diverging ~ 10 million years ago, with a generation time approx 10 years, this equates to average of 60 mutations (base pair substitutions) per generation.

Bacteria with 3 megabases of genetic difference, over 100,000,000 years, average generation time of 1 minute (to keep it simple), equates to average of 0.000000095 mutations per generation.

Why the very significant difference? The difference in size between bacteria and human genomes (.0016) doesn’t seem nearly enough statistically enough to explain that kind of divergent data.

This tends to complicate comparing us vs. chimpanzees to bacteria vs. bacteria:

Hopefully this blog post can help elucidate why the differences in our genomes vs. chimps are indeed produced by mutations when one compares single base pair differences:

There is a difference between describing a capability and a norm. Humans aren’t echolocating mammals. But we have the biological capability to develop the ability. I think that was Dennis’ point.

You might want to read the paper referenced. If that is paywalled try this link. FWIW, E.coli has about the same point mutation rate per base pair per division as humans, roughly within an order of magnitude. For E. coli this means about 0.001 point mutations/generation. However, selection limits the rate of accumulation over time and so the rates of point mutations will be greater than the net rate of variation in a species.

Neutral mutations fix at the rate they’re produced and that is a function of the overall mutation rate and the odds of hitting a selectively neutral site. Compared to E. coli, humans and chimps have a much higher percentage of DNA that is selectively neutral to mutation and consequently seem to have accumulated a greater percentage of mutations per year since these ape lineages split.

Aside: E.coli can replicate every 20 minutes under optimal conditions but inside animal intestines, doubling times are actually closer to a day (or days) – probably related to the transit time through the intestine.

1 Like

Except that the article you linked seems to imply that a lack of change would be one of those “falsifying” features?

I appreciate the link, but… Not to sound overly critical, but I found the the article poorly written and disorganized, and lacking in specific criteria for falsifying evolution, and/or offering criteria that would certainly be disputable. I had trouble even outlining what are the specific falsifying criteria… the author mentions …

1. “Fraudulent claims.”

Except a platypus. But of course that didn’t disprove evolution, so I doubt the author of this article would immediately concede if we ever discovered feathered rabbits. “Wow, evolution surprised us again” would be the response, no? If I believed in evolution, and discovered feathered rabbits, my working assumption would be that we had far more fossils to discover, or a striking de novo gene, or the like, not that evolution was thereby falsified. There’s no scientific reason, assuming evolutionary premises, that it could not have evolved rabbits with feathers, or could so in the future. This doesn’t falsify the macroecolutionary theory of natural selection/variation itself.

2. Deep time

Um, OK. But this is used as a hypothetical contrary to fact, this has nothing to do with actually testing and falsifying the actual biological process itself. I could just as easily observe that if the earth were entirely composed of molten lead, that would be a problem for evolution, too. But I fail to see how this could in any way be used to “falsify” the actual biological theorem itself?

3. No foresight

This one, philosophically, is simply a mess, unusable, and and certainly non-scientific…

Sure he could have. And he could have chosen not to. Philosophically, this proves nothing, and it is certainly no scientific way to falsify and/or prove a theorem, to say, “if I’d designed this, I would have done it differently. Therefore, evolution.”

And if you ever wanted a good example of a “question begging fallacy,” here it is. (Not to mention I am dumbfounded by those who argue for evolution based on how badly we’re designed. Once someone designs a human from scratch which is demonstrably better than the current design of Homo sapiens, only then would I entertain such an argument.)

Any chance you could point me elsewhere, and/or give me any other thoughts about how one could legitimately, in a scientific/empiric manner, actually test and potentially falsify macroevolutionary theory?

Furthermore not all organisms evolve at the same rate. So you cannot compare bacteria to mammals and ask why the bacteria don’t evolve in radical ways after a million generations as mammals have. Evolution accelerates as species evolve more sophisticated ways of introducing variation into their genome. And the evolution from prokaryotes to eukaryotes took billions of years. Bacteria are asexual (simply divide in to exact replicas) and thus depend on either UV radiation or exchanges with other bacteria. UV radiation is very random and exchanges with other bacteria would spoil the isolation requirements of an experimental system.

You might even consider the evolution of prokaryote to eukaryote to be something of a miracle or, at the very least, poorly understood for a number of reasons such as very different earth conditions and greater difficulty in obtaining fossil evidence. Many believe there is evidence that the eukaryote evolved from a symbiotic relationship between different prokaryotic organisms because chloroplasts and mitochondria have some of the characteristics of a prokaryotic cell. We have little doubt that the principles of evolution are capable of producing such an evolutionary result but no idea how likely such a development may have been.

“Let your conversation be always full of grace, seasoned with salt, so that you may know how to answer everyone.” -Colossians 4:6

This is a place for gracious dialogue about science and faith. Please read our FAQ/Guidelines before posting.