Bananas are radioactive—But they aren't a good way to explain radiation exposure


Just look at that radioactive banana. There's nothing special about it or where it was grown. All bananas are radioactive, because all bananas contain the radioactive isotope Potassium-40. In fact, a lot of things you might not suspect of being radioactive are, including Brazil nuts, and your own body. And this fact is sometimes used to downplay the impact of exposure to radiation via medical treatments or accidental intake.

A post by nexusheli on the Submitterator turned me on to the idea of the Banana Equivalent Dose—a way of putting radiation exposure into context by comparing intake from, say, local milk just after the near-disaster at Three Mile Island, to intake from a normal, workaday banana. Wikipedia explains the point of this exercise:

The banana equivalent dose is the radiation exposure received by eating a single banana. Radiation leaks from nuclear plants are often measured in extraordinarily small units (the picocurie, a millionth of a millionth of a curie, is typical). By comparing the exposure from these events to a banana equivalent dose, a more realistic assessment of the actual risk can sometimes be obtained.

This isn't just about pro-nuclear propaganda. "Picocurie" is one of those words that really doesn't mean anything to lay people. Confusing units of measurement, when combined with the word "nuclear" can, understandably, freak people out. And, depending on the situation, there isn't always cause for said freak out. Having a way of explaining what picocurie means, in the context of everyday, normal, safe exposures, really is a useful tool for putting radioactivity into a context the public can understand.

But, the Banana Equivalent Dose probably isn't the best way to do that.

The problem is that this system implies that all radioisotopes are created equal—That there's no difference between 520 picocuries of Potassium-40 and a similar intake of, say, radioactive iodine. And that simply isn't true. I contacted Geoff Meggitt—a retired health physicist, and former editor of the Journal of Radiological Protection—to find out more.

Meggitt worked for the United Kingdom Atomic Energy Authority and its later commercial offshoots for 25 years. He says there's an enormous variation in the risks associated with swallowing the same amount of different radioactive materials—and even some difference between the same dose, of the same material, but in different chemical forms.

It all depends on two factors:
1)The physical characteristics of the radioactivity—i.e, What's its half-life? Is the radiation emitted alpha, beta or gamma?

2) The way the the radioactivity travels around and is taken up by the body—i.e., How much is absorbed by the blood stream? What tissues does this specific isotope tend to accumulate in?

The Potassium-40 in bananas is a particularly poor model isotope to use, Meggitt says, because the potassium content of our bodies seems to be under homeostatic control. When you eat a banana, your body's level of Potassium-40 doesn't increase. You just get rid of some excess Potassium-40. The net dose of a banana is zero.

And that's the difference between a useful educational tool and propaganda. (And I say this as somebody who is emphatically not against nuclear energy.) Bananas aren't really going to give anyone "a more realistic assessment of actual risk", they're just going to further distort the picture.

Geoff Meggitt has written a book about the history of radiation and protection. It's called Taming the Rays, and is available on

Image: Some rights reserved by Jason Gulledge


  1. I really like the idea of a banana equivalent dose, or some way of making units realistic. Aside from mode of intake (through skin, ingestion, etc) wouldn’t it be possible to apply a correction factor to the b.e.q. to account for differences of the first kind (half-life, radiative rays) to make the b.e.q. more useful?

    By the way, this topic is closely related to my pet theory of a contributing factor to the obesity epidemic in the US. All our nutritional information (serving size, amount of fat, etc) are in SI units, which are meaningless to people who weigh themselves in pounds.

    1. Looking at the US unit system it seems as if it were more effective to use metric units because that would eliminate a lot of the unit conversions that are inherent to the US system, where, essentially, an “ounce” can mean 1/16th of a pound, 1/12th of a troy pound or the volume of slightly more than an ounce of water. But maybe that’s just my european view of the situation ;-)

    2. wouldn’t it be possible to apply a correction factor to the b.e.q. to account for differences of the first kind (half-life, radiative rays) to make the b.e.q. more useful?

      If I read the article correctly, no. The problem is that the banana has no net effect on the body at all. As Maggie’s article says, “The net dose of a banana is zero”. So there’s no amount of bananas a person could eat that would result in an equivalent dose of radiation from other sources (except for sources that also add no radiation at all).

      1. My point is that it would be useful (and possible) to have a unit of radiative dose that is separate from how your body deals with the radiation. So while your body can remove excess potassium-40, you are still receiving 1 b.e.q. every time you eat a banana.

  2. A good point. What do you think is a good tool to explain radiation dose, other than bananas? I worked a summer at particle accelerator in 2000 or so. Our radiation safety training, in addition to learning how to use different kinds of dosimeters, geiger counters, warning equipment, etc. included a ‘contextualizing the risk’ section. First they talked about comparable dose (they used bananas, cosmic ray exposure from airplane flights, naturally occurring radon, and medical x-rays as references). Then they pointed out their safety standards were much tougher than the government’s, maybe by an order of magnitude. Then they showed the risk of getting hurt doing various other things vs spending time at the lab. Some comparisons seemed disingenuous but others were eye-opening. And some were just confusing. Driving twenty minutes in a car, or eating two table spoons of peanut butter, pose the same risk of injury as working at the plant for multiple years.

    Not sure how they calculated that. :-)

  3. Iodine 131 is a particularly good counter example. It is a common decay product, it concentrates heavily in the thyroid, and is a gamma and strong beta emitter.

  4. Here’s something else that’s radioactive that most people don’t know about: tobacco grown in the United States. See for some official information (I never thought I’d see the day when the US Government would actually deal with this information directly).

    It gets even worse though – notice how the tobacco grown in the US is radioactive because of the choice of fertilizer according to even the EPA. Note how the EPA and other Government entities are helping to control the risk: measures dedicated to reducing smoking. Why aren’t there laws passed to restrict the type of fertilizer used in tobacco growing to ones that don’t involve radioactive compounds?

    What’s worse is that the US has higher lung cancer rates of smoking than other nations, even nations where smoking is more prevalent. See and

    Smoking is not good for you but if we are going to argue that it needs to be controlled because it causes lung cancer perhaps we should investigate the case that it is such a nasty carcinogen as a byproduct of the manufacturing process and one that can be resolved at that. Of course the Government likes taxes (what does the Federal government do with the tax money anyway? Does it go to help private health care?), people hate the smell of smoking, and smokers are in the minority, so those two groups nicely gang up against the smokers.

    1. GreenJello,

      We aren’t paid, nor do we pay anybody, for Submitterator items. The site, itself, is sponsored by AmEx, but that just means they supplied the funds we needed to set the site up in exchange for having their name on it.

      The actual submissions come from readers who have made or found something cool they think we need to know about. It’s no different than “Submit an Idea”, just in a different format that’s easier for us and you to use.

  5. The largest problem about the whole topic is the absolute lack of trust combined with a perfect unwillingness to educate oneself.

    People can openly say something like “radioactivity is dangerous, that’s all I need to know”, without anyone realizing that this is bogus. Because this only makes sense exactly until you run into the next activity that is in any way dangerous.

    How about going to war? That’s something the US does about twice a decade. Hunting Osama bin Laden and Saddam Hussein alone cost a lot more lives than all nuclear accidents that happened so far ever will, even if you ask Greenpeace. (A reliable source for collecting and adding up only the scariest numbers they can find, regardless of serious and objective criticism. – Their number is 200,000 in this case.)

    Criticism on the openly admitted basis of purposeful ignorance and unwillingness to learn (and only that kind!) should finally get a well deserved, sharply focussed(!), treatment of ridicule.

    There can be no trust on the basis of ignorance.

    The knowledge is out there, it’s not hidden, it’s not new science and it’s in fact a lot less complicated than epidemiology, economics, let alone politics.

  6. I’m working on the Waste Treatment Plant being built in Hanford, WA. I think the last radiation safety course I took used chest x-rays to contextualize the amount of radiation a worker can expect to encounter when doing certain tasks.

    However, I think all of these examples are fairly useless. Telling me that changing a HEPA filter is equivalent to 10 chest x-rays simply makes me ask “Well, how much danger is presented by a chest x-ray?” How many cells are mutated by the ionizing radiation? How many of those are likely to be cancerous? Without that information, the contextualization is useless (it’s even worse with bananas which have associated idea of good health).

    1. “Well, how much danger is presented by a chest x-ray?”

      More than six in a year is believed to represent an unnecessarily high risk.

      Keep in mind, though, that the guideline is trying to balance likelihood of finding a dangerous condition (lung cancer, broken sternum, collapsed lung) against likelihood of causing disease (lung, skin or bone cancer, blood diseases, etc.) It’s understood that properly administered chest Xrays can be extremely beneficial, and that having a chest Xray every day is unlikely to provide a doctor’s ability to diagnose your problems.

      So, while chest Xray equivalents are better than bananna equivalents, it’s still a pretty lame measure. Your random exposure in the lab is not going to ever result in a beneficial medical image.

    2. Nothing like forgetting to post half your message is there. *sigh*

      I’d love to hear more about your job lasttide. This year the Waste Treatment Plant construction site is off limits but I’ve seen it a few times during past Hanford tours.

      1. Unfortunately, MadRat, you will need an instrument a lot more sensitive than a pancake-type GM probe to see the tiny amount of activity found in a banana. It might be possible if you could create an artificially low background area. Take a look at the formulas for minimum detectable activity (MDA)or the Lower Limit of Detection (LLD). If your activity is extremely low, then you need high efficiencies and low backgrounds to see it. (Incidentally with the gamma at 1460 keV, I wouldn’t expect to see more than 20 – 25% efficiency)

        1. Excellent comment. Yes, I could use a $2000 scintillation meter for serious accuracy but as you pointed you that still wouldn’t remove the background radiation. The US and Russian governments have rooms, made from pre-WW2 materials, designed to block out background radiation. I imagine with a banana you could just use a box instead of an entire room.

          The point is there’s so little radioactivity in our food that you can’t even be sure you’re actually measuring it. So, everyone, relax and enjoy your banana split for crying out loud!

  7. Peanut Butter is dangerous. (After all, this is all about cancer risk, not radiation sickness.) I’ve long been trying to put radiation risk in terms of colon cancer caused by …peanut butter sandwiches. PBJ EQUIVALENT DOSE. See radiation risk

    40 tablespoons of peanut butter (20 sandwiches) is equivalent to 0.01 Rem, equals smoking 1.4 cigarettes, equals driving 40mi, equals 6min in a canoe, equals chest x-rays.

    Cross country airline travel exposes us to quite a bit of radiation. Cancer risk from each hour of flying is equivalent to eating 1.4 PBJs, or 30mi of highway driving.

  8. The Potassium-40 in bananas is a particularly poor model isotope to use, Meggitt says, because the potassium content of our bodies seems to be under homeostatic control. When you eat a banana, your body’s level of Potassium-40 doesn’t increase. You just get rid of some excess Potassium-40.

    Uh… okay, but what if you like to smush up bananas and spread them all over your skin and roll around in banana mush and plastic sheeting? Like… um… say four or five nights a week? Would that be, uh, very dangerous? PLEASE RESPOND IMMEDIATELY!

  9. Using a different unit helps, health physics already has one for radiation absorbed by humans, rem. It’s a unit of dose, not just activity (like curies).
    Normal background radiation is ~360mrem/year, so about a mrem/day.
    Knowing how many mrem we receive in a day helps us understand the dose of other things like x-rays (chest – 11mrem, panoramic dental – 30mrem).
    I used to describe dose to people in days of background radiation.

  10. Oh no — don’t start abbreviating the “banana equivalent dose” as “b.e.q.”! Way too close to becquerel (Bq), yet another unit of radiation.

    From Wikipedia:

    “The becquerel (symbol Bq) is the SI derived unit of radioactivity. One Bq is defined as the activity of a quantity of radioactive material in which one nucleus decays per second. The Bq unit is therefore equivalent to s−1. The becquerel is named for Henri Becquerel, who shared a Nobel Prize with Pierre and Marie Curie for their work in discovering radioactivity.”

  11. EXACTLY, cjeam read my mind. REM and Sievert units account for normalization across radiation type and exposure. The open question that /should/ be talked about and where folks /should/ want more education is the concept of linear / no threshold somatic and genetic effects.

    We know lots of dose is bad. Ask Louis Slotin.

    We know moderate amounts of dose is bad. Ask Russian Emergency Response Workers.

    But where does “bad” start for low dose? Is every dose bad? That is the assumption of the linear / no threshold model. Take data points of risk of effect plotted against dose and force fit that line back to zero. There has not been enough study to determine if low dose has no effect (ie, the dose intercept is not 0.) There has not been enough study to determine if low dose is good (“priming the immune system” argument). There’s a lot to think about.

    Making radiation fun with goofy “banana metrics” is probably not the best idea. Keeping radiation as the bogeyman of the modern age it’s been for the last 70 years is also probably not the smartest thing in the world.

    1. “But where does “bad” start for low dose? Is every dose bad?”

      The evidence for hormesis (that the dose response follows a j-shaped curve that for low doses dips below the risk posed at zero dose before rising back up) seems pretty strong. For instance there is a correlation between natural radon emissions and lung cancer rates – and that correlation is negative. See for a collection of charts of radiation vs. mortality from many studies.

      A review of the literature is here:

      Some highlights:
      According to UNSCEAR report (1994), among A-bomb survivors from Hiroshimaand Nagazaki who received doses lower than 200 mSv, there was no increase in the number of total cancer death. Mortality caused by leukemia was evenlower in this population at doses below 100 mSv than age-matched controlcohorts.

      Mifune (1992) (Mifune et al. 1992) and his co-workers indicated that in a spa area (Misasa), with an average indoor radon level of 35 Bq/m3, the lung cancer incidence was about 50% of that in a low-level radon region. Their results also showed that in the above mentioned high background radiation area, the mortality rate caused by all types of cancer was 37% lower

      In a very large scale study in U.S.A, it was found that the mortality rate due to all malignancies was lower in states with higher annual radiation dose (Frigerio 1976).

      In a large scale Chinese study, it was showed that the mortality rate due to cancer was lower in an area with a relatively high background radiation (74,000 people), while the control group (78,000 people) who lived in an area with low background radiation had a higher rate of mortality (Wei L 1990).

  12. In the end, the only reasonable way is to quit the baby-talk and talk about the way it is.

    Natural radiation is, on average, about 1 REM in about 3 years or 0.01Sv. There are large areas that have twice as much, others have half as much.

    One of the big problems in saying how exactly those doses affect health or growth of cancers is that there are a lot of other factors that are much worse than triple the amount of background radiation and difficult to account for. Like smoking, living in a chemically contaminated area or local outbreaks of retroviruses … (those guys are designed by nature to meddle with your DNA and they don’t care that much whether you come down with cancer or not – and we only know those viruses that actually cause symptoms)

    That’s not to belittle the problem or say it doesn’t exist. But there obviously is a bit of a difference between reality and the public notion that any amount of radiation, no matter how small, will cause cancer rates to sky-rocket.

  13. This might be a distortion but it isn’t as bad a distortion as calculating radioactivity in terms of “atomicity” (number of radioactive atoms). That enabled an alleged scientist to claim that a ton of depleted uranium has the equivalent of 100 Hiroshima bombs.

  14. One thing seems to be missing from most of this discussion: Radiation isn’t bricks (kg), it’s watt-hours.

    The article says “The net dose of a banana is zero”, but this is no more true for potassium-40 than it is for carbon-14.

    Even if it passes on through the body, you’re exposed to whatever amount of radiation it emits for that amount of time.

    I wouldn’t worry a lot, though. Potassium-40 has a half-life on the order of 10e9 years, as opposed to Cabon-14’s 40 years.

  15. from the perspective of the question of homeostasis of potassium isotope balance, the banana does not represent a nonzero amount of extra potassium 40. it’s like saying ‘the water is too limey, let’s pour more of the same water in to make it less limey’. the potassium in a banana has an intrinsic toxicity also, if you were to ingest more than a certain amount of certain forms of potassium that is from such things as pH and specific chemical affinities. ah yes, differentials. in my opinion to speak of equivalency with a substance i’d say (milli)grams of uranium ore. the thing about radioactivity is there is numerous factors as mentioned, exposure to alpha and beta and gamma and all that jazz is different per substance, the affinity of the radioactive substance with specific organs, eg iodine and the thyroid gland, and how it alters their chemistry in critical ways to cause mortality. the different amount of protons and neutrons in a radioactive substance alters its reaction tendencies as well as the next stage of the radioactive decay which means a bunch of nasty things end up forming, lead being one of them obviously.

  16. I have a laboratory quality Geiger counter (digital with a pancake detector) and placed it on a banana for 10 minutes. I couldn’t detect radiation above background. It could be that the peal blocks some of the radiation (not impossible) but I didn’t want to put a freshly pealed banana on my nice, clean Geiger counter.

  17. When I made my comments to Maggie I certainly wasn’t too smart in what I said about banana dose. It seemed to me that if you didn’t eat the banana, you got the potassium because of the homeostasis (and the dose) from some other sources. Plus, I thought the homeostatic effect worked before the potasssium got out of the bloodstream and it was pretty quickly dumped with negligible dose. Opinion I now find seems to be that it is distributed around the body and lingers for about 20 days. The dose is still very small and I make it about 0.1 microSv or 10 microrem.

    This makes annual natural background equivalent to about 20000 bananas a year. So here’s my new problem with the BED. If somebody told me that eating a banana would increase my radiation dose by about 1/20000 of what I would (inevitably)get in a year (so say 1/2000000 of my lifetime background dose) it would make some sense to me. It would tell me that bananas are radiologically speaking pretty safe. On the other hand if I was told my annual background dose is equivalent to eating 20000 bananas it wouldn’t help me much.

  18. This article has a number of problems. As just one example, the differing penetration power of beta and gamma are irrelevent once the material is biologically absorbed– the Q Factor of both is unity.

    I won’t go into the other problems, but if you don’t like bananas as a comparison, what about Brazil nuts? They’re more radioactive than bananas (some batches up to 3X more so, depending where grown), and because the radiosource in this case is radium, rather than potassium, its not under homeostatic control.

    There are countless other foods that are radioactive. The “Banana Equivalent Dose” concept isn’t about bananas; its about understanding that radiation is extremely prevalent around us, and our ability to detect it far, far exceeds its ability to harm us.

  19. “there’s so little radioactivity in our food that you can’t even be sure you’re actually measuring it.”

    Sorry, but that couldn’t be more wrong. For the average person, more than 10% of your annual background dose comes from food source– 30 to 40 millirems worth. When you have nitwits suing nuclear plants for exposures in the picorem level (one billionth as much), that’s an important consideration to remember.

  20. There have been times I’ve been wrong while making comments on BoingBoing, so I could be wrong now. Here’s the way I’m thinking:

    Radiation is very random and measuring it’s complicated. What kind of radiation is it, how much energy does it have and what are you using to measure it? How far away is the source of the radiation and for how much time did you measure it? What other sources of radiation were detected? Some materials, even air, can block radiation and other materials convert it from one type of radiation to another, so what was between the detector and the source?

    Even more complicated is the yearly dose. What kind of building do you live in? What elevation do you live at? What’s in the dirt were you live? What do you work with that exposes you to radiation? How many flights did you take and how long were they? How many medical x-rays did you get and what kind where they? The list goes on and on. The only way to be sure of the yearly dose is to wear a dosimeter for a year, without ever removing it; anything else is just a estimate.

    Instead of trying to answer all those questions let’s over-simplify. Let’s say one millirem per hour (mR/h) equals 240 counts per minute (cp/m), there are 365 days in a year and, just to make it easy, the yearly dose is 365 mR/y. There are 24 hours in a day so 1 mR/d divided by 24 hours gives you about 0.042 mR/h. If you multiply 0.042 mR/h by 240 counts per minute you get about 10 counts per minute. If our food gives us 10% of our yearly radiation dose, then how do you measure which of those 10 counts per minute comes from food? Wouldn’t you agree that would be hard to do even while ignoring all the other complications?

  21. Rat, measuring a radiation dose is quite simple, even without an integrating counter. There are only three basic types: alpha, beta, and gamma…and there is a precise formula for each for converting from becquerels into sieverts…a conversion that accounts their varying effect on human tissue. When you have multiple sources, you measure each independently, then sum. What’s so hard to understand about that?

    The only difficult portion is the local variance — the “average” dose of 360mrems can be well over 1,000 in some areas. However for the purposes of THIS discussion, that’s beside the point. The contribution from food sources is one of the most static factors; nearly all the variance comes from radon and cosmic rays, which depend on soil composition and altitude, respectively.

  22. I didn’t read every comment so it is possible that this has been mentioned. While I understand there is already 40-K in your body, when you eat a banana there is roughly a 24 hour period from when you consume the 40-K and excrement the “extra.” So yes, eating a banana SLIGHTLY increases your received dose. Speaking in terms of EFFECTIVE dose with respect to decay mode, it doesn’t matter greatly since 40-K decays via beta, gamma, AND alpha (mostly beta).

  23. Anon with post #35 is bang on. The BED is valid because it is used to compare equivalent dose. Not general radioactivity.
    It doesn’t compare curies (or bequerels), it compares sieverts (or rem). A sievert from K-40 is the same as a sievert form any other radionuclide.

  24. Maggie, thanks for putting the record straight on all this banana bullshit. It all seems to have started years ago by some poorly thought out and technically incorrect comment from some learned body in the States.

    #35 sums it up well. the K-40 will have a very small but finite effect because it is inside the body for about 24h. Even though it is not assimilated.

    However, this effect will be many orders of magnitude smaller than what is being quoted by the banana bunch who seem to have 2c worth of knowlege about physics and absolutely zero in biology.

    Interesting to see how eagerly this sort of crap is applauded and repeated by pro-nuke propagandists.

    [Reading the link in #35 it is clear that there is NO alpha from K-40 . That is 20x more dangerous than beta or gamma so just as well.]

  25. @ banana

    Hold on here, it sounds like you’ve just agreed with the #35, yet are leaping to a conclusion that is the precise opposite of the actual discussion. Which was quite intelligent, and largely free of political spin, might I add.

    A banana equivalent Dose is measured in microsieverts (μSv), an actual unit of ‘dose’. As discussed, the important thing is to realize that sieverts (and rem) represent a relative quantification of the physiological effects of radiation, based on a number of physical factors (e.g. type of decay particle).

    Recapping: 0.1 μSv is the dose you receive from bananas.* 0.34 μSv/hr is the average background dose for Americans. As of right now, the dose in Tokyo is approximately 0.2 μSv/hr. ( – 100 CPM = 1 μSv/hr).

    While saying that simply living for a year is the radioactive equivalent of eating 30,000 bananas is indeed a rather meaningless, arbitrary, and pretty stupid comparison, that’s actually the entire point. A banana represents an utterly insignificant, vanishingly tiny amount of radiation. But guess what: for the majority of people in Japan, so does Fukushima.

    Your odd snap conclusion about pro-nuclear propagandists seems to indicate that you believe that radiation from nuclear power represents a threat to public safety. And yet, given that the hourly dose in Tokyo right now is just over half that of the average dose in America in the midst of what is absolutely an extremely serious nuclear calamity – the 2nd most serious in history, and by a large margin! – brought on by truly extraordinary circumstances, wouldn’t you say that the facts seem to indicate no widespread threat whatsoever?

    *I’m assuming this dose takes into account the homeostatic properties of potassium-40, but maybe it doesn’t – for a more apt comparison, feel free to use any of the other ‘radioactive’ foods out there, like Brazil nuts. Or maybe the dose from radon gas you receive in your basement due to the concrete cinderblocks.

  26. I would like to jump in the discussion here because it seems like people are comparing two completely different things.

    There is a huge difference between radiation and radioactive materials. You can suffer a lethal dose of radiation (especially gamma rays) with actually coming into contact with a single radionuclide. You take a good dose of radation flying in a plane or getting an X-ray, but you are not exposed to any radioactive materials.

    On the other hand, you could come into contact with radioactive material not get enough of a dose to couple any problem what so ever. I worked my way through college packaging nuclear medicine. People (including myself on one occasion) would sometimes get something on their hands and have to be decontaminated (which for me simple involved a thorough washing on my hands and checking to makes sure I did not have it anywhere else on my body or clothing).

    The banana equivalent dose is just that – the equivalent to a dose of radiation. Not the equivalent to ingesting an equal amount of some other isotope. It is not a way of saying that K-40 is the same as I-131 or Cs-137. It was invented as way of saying two things: 1) you are exposed to radiation and radioactive materials in everday life and 2) here is a way to measure it using a common object.

    I think it is excellent for that purpose so long as you explain what you are comparing.

Comments are closed.