Nature Has A Formula That Tells Us When It's Time To Die

January 23, 2013 4:29 AM

Here's the surprise: There is a mathematical formula which says if you tell me how big something is, I can tell you — with some variation, but not a lot — how long it will live.

Yunfun Tan illustrates the heartbeat of mother nature in this post on NPR
posted by rebent (33 comments total)

Hmm. Please show your workings for the tortoise.
posted by jaduncan at 4:36 AM on January 23, 2013 [1 favorite]

Yes, it's clear from the graph that 'some variation' is a couple of orders of magnitude.
posted by edd at 4:38 AM on January 23, 2013 [4 favorites]

What edd said. If someone plotted this on a linear scale, you'd write the whole thing off as absolute bullshit, not some eerily prescient formula. Log-log plots are really good at two things. They linearized data of the form y=Ax^B and they make absolute shit look really good in your graph, because people are phenomenally bad at thinking in log scales.
posted by Kid Charlemagne at 4:58 AM on January 23, 2013 [8 favorites]

I can tell you — with some variation, but not a lot — how long it will live

I can, too. Generally up until the time that humans decide either a) you are tasty or b) provide erections. After that, it's a short slope down.
posted by GenjiandProust at 5:10 AM on January 23, 2013 [7 favorites]

Out of scientific curiosity, is there something that combines a and b?
posted by KMB at 5:18 AM on January 23, 2013 [2 favorites]

It's true. Last Saturday a never-before-known ugly weasel-like creature was discovered in Fresh Kills and not only was it really tasty, but it gave you a two hour hard-as-steel boner AND shortened the refractory period to ten seconds. Just try to find one now.
posted by seanmpuckett at 5:19 AM on January 23, 2013 [2 favorites]

is there something that combines a and b?

Your first clue is an erection that lasts more than four hours.
posted by localroger at 5:24 AM on January 23, 2013 [2 favorites]

Wait, seriously? They had a scatter with three big clusters in log-log and they drew a line through the three centroids and they were all POWER LAW???

I should read the paper, it's probably better than my caricature of it.
posted by escabeche at 5:38 AM on January 23, 2013 [3 favorites]

I saw an interesting article a couple days ago, but can't find it now. Basically, it says that the more cells you have, the more likely it is that you'll get cancer, so the more useful cancer-suppression genes become. The energy required to run the extra cellular repair machinery is somewhat costly, so creatures that correctly balance the amount of cancer suppression against their number of cells will produce the most offspring.

That was as far as the article went, but then gluing in some stuff I've learned elsewhere: another strategy for preventing cancer seems to be aging. If a given cancer cell doesn't also mutate to fix its own telomeres, it won't produce enough generations of itself to be truly dangerous before they all die out. Obviously, this doesn't always work, as some cancers do make themselves immortal, but it helps a lot.

So, it looks to me like age and size and cancer risk are all closely linked. If you're big and have lots of cells, you need strong tumor suppression. If you have that, I presume you don't need to age as quickly, as the survival advantages of long experience probably then outweigh the survival advantages of early cell death to prevent cancer.

Small creatures tend toward being highly efficient, not wasting energy on cellular repair, and thus dying very quickly to avoid cancer. Instead of paying in calories to protect themselves, they pay in lifespan. High cost to the individual, strong benefit to the species.
posted by Malor at 5:40 AM on January 23, 2013 [4 favorites]

Dogs seem to be doing this wrong.
posted by localroger at 6:11 AM on January 23, 2013 [1 favorite]

I just *knew* this was going to be about inference of Dr. Everything's Super-Grand-Unifying Laws of Life using log-log plots!

Of course, there is this Science article from a few years back showing why log-log plots may be more bollocks than dog's bollocks in many cases...
posted by aeolicus at 6:17 AM on January 23, 2013 [1 favorite]

Parrots? Jellyfish?
posted by Decani at 6:37 AM on January 23, 2013 [2 favorites]

i fulfill both criteria, yes

Dude, you are advertising yourself for extinction!!!!!
posted by GenjiandProust at 6:39 AM on January 23, 2013 [1 favorite]

See also Stephen Jay Gould's "One Standard Lifespan" ("During an average life, all mammals have nearly the same number of breaths and heartbeats").
posted by yz at 7:00 AM on January 23, 2013 [2 favorites]

Cosma Shalizi's So You Think You Have a Power Law remains an excellent counter-balance to "look I plotted it log-log and it's sorta straight!"

Key point: "Lots of distributions give you straight-ish lines on a log-log plot. True, a Gaussian or a Poisson won't, but lots of other things will. Don't even begin to talk to me about log-log plots which you claim are 'piecewise linear'. "
posted by PMdixon at 7:25 AM on January 23, 2013 [5 favorites]

So...elephants' average lifespan is something like 400 years, and parrots live about 15. Good to know.

Thanks, Science!
posted by etc. at 7:27 AM on January 23, 2013 [1 favorite]

There's another reason why animals have to be more efficient with increased size. As animals get bigger, their volume (x^3) increases much faster than surface area (x^2), which means it becomes increasingly difficult to shed excess heat generated by cellular processes. So while larger animals have higher metabolic rates compared to smaller animals, their metabolic rates are proportionately much smaller per unit of mass.
posted by dephlogisticated at 7:31 AM on January 23, 2013 [2 favorites]

The fatter I get, the longer I live? Cool!
posted by dances_with_sneetches at 7:31 AM on January 23, 2013

Out of scientific curiosity, is there something that combines a and b?
posted by KMB at 5:18 AM on January 23 [1 favorite +] [!]

Chocolate and coffee. Or were you looking for an animal?
posted by Stagger Lee at 7:38 AM on January 23, 2013

argh yes everyone read that fracking Cosma Shalizi post like yesterday.
posted by en forme de poire at 8:29 AM on January 23, 2013

Ming
posted by bonobothegreat at 8:38 AM on January 23, 2013 [1 favorite]

I think you get more accurate results looking at heart beats / lifetime (i.e. number of heartbeats remains pretty constant). At least for mammals.
posted by YAMWAK at 9:31 AM on January 23, 2013

Please show your workings for the tortoise.

Tortoises are bigger on the inside.
posted by yoink at 9:45 AM on January 23, 2013 [5 favorites]

Nah - tortoises work fine on average - just remember that they get bigger the further down you go.
posted by YAMWAK at 9:52 AM on January 23, 2013 [1 favorite]

I think you get more accurate results looking at heart beats / lifetime (i.e. number of heartbeats remains pretty constant). At least for mammals.

Metabolic rate being inversely correlated with longevity is an appealing idea (those overachievers are literally burning away their lives! Turtles may live a long time, but they're penalized by being slow!), but alas, there are those pesky birds and bats with lifespans over 3x that of similarly sized mammals, despite their supercharged metabolisms.
posted by Pyry at 10:03 AM on January 23, 2013 [3 favorites]

This article was just an excuse to justify using those plant animations.
posted by davejay at 3:10 PM on January 23, 2013 [1 favorite]

This article is really missing the point of this body of research and so, understandably, are you guys in your dismissal of it. The reason why do plots of body size and life span is because it gives you a null metabolic expectation; how long is a given thing expected to live under basic metabolic/evolutionary constraints? This is a really important thing to do, because it then allows you to ask what gives rise to the (considerable) variance seen around that line. Someone above mentioned that bats trend very differently than the rest of mammals. Doing this sort of plot is a prerequisite for figuring out how much they deviate from expectations, which is a prerequisite for figuring out why. This, in turn, will give us insight into why such trends exist, and maybe ultimately into the cap on our human lifespans. Or, better yet, some basic scientific insight. So it's useful, cool science, just the claim in the title is a more of the ridiculous journalism you see everywhere. Keep in mind that the first weather forecasts, too, were overwhelmed by variance, but they gave a jumping-off point to learning a lot of interesting things.
posted by Buckt at 6:04 PM on January 23, 2013

buckt, the reason people are dismissing it is that it is such a vague correlation that the causality it implies is almost certainly meaningless if it exists at all. It's like saying larger things travel further. Yeah, sure, whales migrate all the way around the world but you know what? Some hummingbirds and butterflies get almost as far. Parrots and pigeons occupy the same ecological niche; they eat the same things and live in the same environment, but pigeons live fast, die young, and reproduce like rabbits to compensate. Amazon parrots, about the same size, live as long as we do and reproduce about as often. Body size is a factor that influences life expectancy but it's really only one small variable and probably has more to do with group selection for population stability than individual metabolism.
posted by localroger at 6:37 PM on January 23, 2013

I think my weather analogy fits very nicely as a counter to what you just said. A single factor doesn't explain life span - that's clear. But if we want to figure out what does predict life span, we need to incorporate relevant variables into the model and, for the time being, body size is a pretty good one. It gives us a null expectation and when organisms depart from it, we can go looking for our second and third components of the model. Also, we both must know that stating "it is such a vague correlation that the causality it implies is almost certainly meaningless if it exists at all" is ridiculous - body size variation within smaller clades, species, and (if we're talking about ants...) even among sisters corresponds extremely well with life span. It's only when you do these super large scale analyses do the trends weaken, but, again, given the diversity of life on earth, who would expect it to be any different?
posted by Buckt at 8:27 PM on January 23, 2013

Fine - let's say that plotting body size against average expected life expectancy hints that the two are correlated. You're not going to convince too many people by waving a log plot around and saying "Look! It's obvious!".

Now do a little statistical analysis and give me a correlation coefficient with a decent p value, breaking it down by category and with the group as a whole and maybe it'll be more convincing.
posted by YAMWAK at 10:40 PM on January 23, 2013 [1 favorite]

It's only when you do these super large scale analyses do the trends weaken

Is it a "super large scale analysis" to bring up the fact that Canis Familiaris trends almost exactly opposite to this across a wide range of breeds?

The OP suggests that there is a direct causal influence between body size and life span -- "Nature has a formula." The thing is, that's the kind of assertion that is easily knocked down by a single counterexample. If Nature has a formula then she needs to buy a new calculator. There are quite a few very dramatic counterexamples. And while yes, there is a very vague correlation between body size and life span, that correlation does not imply causality. There are certain broad influences that make live fast / reproduce a lot a better strategy for small species but wherever other strategies are workable, some species find them.
posted by localroger at 5:36 AM on January 24, 2013

I think you're arguing something different than I am; I said in my first comment that the article is a misrepresentation. Your last statement is the only point I was trying to make; there are evolutionary reasons why these trends exist, and studying the places where organisms have found alternative strategies and broken the trends is useful. And your correlation/causation comments are also a bit of an aside. I don't think anyone claims that size causes life span; the article mentions the confounding factors, such as cell efficiency or metabolic rate, that are the true candidates for causation. This thread is just a lot of misdirected anger, to me, and, as I said a moment ago, that's partially the fault of the article.
posted by Buckt at 10:07 AM on January 24, 2013

Yeah Buckt I guess we don't disagree, but the article is pretending to draw something really profound out of a half and half mix of innumeracy and woo.
posted by localroger at 11:32 AM on January 24, 2013