How often do bird species hybridize?

March 5, 2020 • 10:15 am

There are many reasons why we want to know how often distinct species hybridize, i.e., form individuals resulting from the mating of a male from one species with a female from a different species. For one thing, if this kind of mixing was very frequent, it would be hard to recognize distinct species as the hybrids would form a continuum between the parents. This isn’t a problem, as species (most notably in birds, as documented in my book Speciation with Allen Orr) remain pretty distinct. But if species remain distinct despite even pervasive hybridization, as seems to happen in some groups like ducks, this is evidence that the hybrids themselves are not mating back to the parental species and blurring species boundaries. And if that’s the case, then we can ask why hybrids are effectively sterile. This could be because they are physiologically sterile, like mules, or “behaviorally sterile”: hybrids could be capable of having offspring but might have the wrong appearance or behavior to attract mates, and so would remain unmated.

But hybridization can have other evolutionary effects. It can, for example, act as a form of “mutation”: if the hybrids are fertile, there’s a chance of genes being moved from one species to another, which could then be acted on by natural selection. This phenomenon, called “adaptive introgression”, is fairly well documented in our own species: “modern” Homo sapiens, for instance, shows several genes that came from Denisovans or Neanderthals (some, like me, consider these subspecies), and were probably driven to high frequency by natural selection. This paper gives a lot of examples of adaptive introgression between more well-demarcated species.

Birds are especially good candidates for estimating rates of hybridization, as they are widely observed, there’s a whole “citizen science” project (“eBird”) in which birders and bird lovers send in records of millions of birds, and hybrids are often easily recognized.

But up to now we had very little idea of how often hybridization occurs. The only estimate, and it’s not a great or systematic one, was one made by Ernst Mayr in 1963, who claimed that he observed only about one hybrid among every 60,000 museum skins he examined: a hybridization rate of 0.00167% (0.0000167). But this low value was a purely off-the-cuff “guesstimate”.

Now, in a new paper in Evolution, three researchers used eBird data to get a better estimate of how often bird species produce hybrids. You can access the paper for free by clicking on the link below, the pdf is here, and the reference is at the bottom. It turns out that hybridization, as estimated by citizen scientists, remains low—about the same order of magnitude as Mayr’s earlier estimate.

First, how do you know when a bird is a hybrid? It’s usually done by observing a weird bird that combines the morphological characteristics of two species who have the opportunity to hybridize. (It could be verified genetically, but that’s hard to do with wild birds.) However, the intermediacy, which is rare, is usually verified by experts, which is how hybrids find their way into the eBird database.

There’s a whole site devoted to bird hybrids, and that is its two-word name. I’ll show three pictures of putative hybrids from the site. Here, for example, is a picture from Bird Hybrids of a hybrid between a mallard and a black duck:

American Black Duck x Mallard hybrid, Ottawa River, Ottawa (Ontario, Canada), 7th April 2016 – copyright Gordon Johnston (photo ID: 2787)

Here’s a likely hybrid between a Great Blue Heron and a Great Egret:

Great Blue Heron x Great Egret hybrid, Fort de Soto Park, Pinellas County (Florida, USA), 17th August 2016 – copyright Dave Norgate (photo ID: 2976)

And a hybrid between a Greater white-fronted goose and a Canada goose, showing the Canada goose parental species:

Greater White-fronted Goose x Canada Goose hybrid (same bird as in photo ID 1844 above; with Richardson’s Canada Goose hutchinsii), location not given (probably Colorado, USA), 25th December 2013 – copyright Cathy Sheeter (photo ID: 1845)

There are several ways to calculate hybridization rates. The simplest is just the number of hybrid birds found divided by the total number of birds observed. But bird species vary in their numbers, and if a numerous species hybridizes a lot, that could give you an overestimate of how often members of all species produce hybrids. To correct for that, you can calculate a per species hybridization rate: the total number of species that have produced hybrids divided by the total number of species observed. (You can also do this for hybridization rates between bird orders or bird families.) This of course will be higher, because a species is counted as hybridizing even if it produces only a single hybrid. Alternatively, you can take each species, calculate its hybridization rate with all other species, and then average that across all species to get an estimate of how often an average bird species produces hybrids. (That was not done in this paper.)

There are of course errors introduced by using observations from eBird. One is the assumption that hybrids are recognized as hybrids just as often as pure-species birds are recognized as pure. That’s not unlikely given that hybrids are often distinct, but it may lead to underestimates if very similar-looking species, like warblers, produce hybrids that, because the parents are similar, aren’t easily seen as hybrids. Conversely, hybridization may be overestimated because birders may report this exciting event more than once, so the same hybrid is counted multiple times. Time of year also matters, as hybrids are best recognized when the parental species are in their breeding plumage: during the mating season. I won’t go into detail about how the authors dealt with these issues, except to mention that they used location information to rule out hybrids that were reported multiple times. You can read the paper for the caveats and biases.

The methods: the authors used eBird observations reported between January 1, 2010 and December 31, 2018: nine years of data. Observations were restricted to the contiguous U.S. and submissions were scrutinized and vetted by experts.

Here are the results, all of which confirm Mayr in suggesting that bird hybridization is rare.

1.) The overall hybridization rate was calculated from 212,875 hybrids reported among 334,770,194 birds, or a rate of 0.064%. That is forty times the rate that Mayr reported.

2.) The corrected hybridization rate eliminating possible multiple sightings of the same hybrid: 0.076%, about the same as the uncorrected rate.

3.) The hybridization rate eliminating species that were very prone to hybridization. Eliminating the 10 most frequently hybridizing species, mostly ducks and gulls, which hybridize like gangbusters, brought the hybridization rate down to 0.009%, about 5 times higher than Mayr’s rate. It turns out that the families Anatidae (ducks) and Laridae (gulls) contributed 83% of all bird hybrids in the dataset.

4.) The species hybridization rate: 242 species were implicated in forming hybrids out of 1146 species available giving a frequency of 21% of species forming hybrids at all.

5.) The order hybridization rate. There are 25 orders of birds in the U.S.; 16 of these had species involved in at least one hybridization event, giving a rate of 64%. Here’s a plot of the wide variation in hybridization rate among bird orders. You can see that the two orders Anseriformes, which contains ducks, geese, and swans, and Charadriiformes, which includes gulls, comprise the bulk of hybridization among birds.

6.) The family hybridization rate. there are 95 bird families in the U.S., and 35 had species involved in hybridization, giving a rate of 37%.

The upshot. Correcting for multiple reports of hybrids and eliminating the sluttiest species of birds, the researchers got a rate of about 5 hybrids per 60,000 species: five times higher than that of Mayr—but Mayr’s value was unreliable to begin with. This still means, however that bird hybrids are rare. Here are the data shown graphically in the paper, with the hybridization rate going down as the sluttiest species are removed (oy, the mallards!):

Are there any implications beyond this? Yes. First, there are observations by my colleague Trevor Price and his associate Michelle Bouvier that crosses in captivity between members of different genera and even families can produce viable hybrids, at least in the “lab”. The absence of such hybrids from nature means either that prezygotic isolation (mating discrimination and other impediments to gene flow that operate before copulation) is very strong, or that viability of hybrids in nature is much lower than observed in captivity. (It’s probably a combination of both factors, but I suspect that mate discrimination is quite strong in the wild, and can be overcome by forcibly confining birds in captivity.)

Second, the genetic data from birds, particularly mitochondrial DNA, shows that species remain quite distinct, with species-specific DNA sequences as judged from their “bar codes”.  This shows that despite even the low rates of hybridization, hybrids are not putting foreign genes into other species very often.

h/t: Luana


Justyn, N.M., Callaghan, C.T. and Hill, G.E. (2020), Birds rarely hybridize: A citizen science approach to estimating rates of hybridization in the wild. Evolution. Accepted Author Manuscript. doi:10.1111/evo.13943

34 thoughts on “How often do bird species hybridize?

  1. Very interesting. I wonder if Mayr’s estimate was skewed by museums only keeping clearly defined ‘types’, discarding any odd ones?

    1. I doubt that museums would do that very often. Odd looking skins would be potentially interesting as either candidate new species or as evidence of hybridisation or simply of the xtent of variation within a species.
      I know that with respect to butterflies collectors were always very interested in specimens that were ‘aberant’ whether due to mutation or hybridisation. I would imagine the curators of bird collections would be similarly interested in oddities.

    1. I will hypothesize a combination of
      1) commonality
      2) habitual association with other ducks in mixed-species flocks
      3) similarity of appearance of female ducks of all species
      4) habit of male-group forced copulation
      5) male avidity

    2. There are many species closely related to the mallard (some authorities probably consider some or most of them as subspecies of the mallard) in which the male is very similar to the female, just a bit brighter and more crisply plumaged. The Black Duck is an example. Mallard males presumably evolved their colourful plumage in order to be more attractive to females (perhaps mallards were in more dense populations and competition was fiercer among mallard males than in the similar species). Mallards have spread all over the place and now come into contact with the more dull-plumaged species (which evolved from a mallard-like ancestor). Presumably the females of these species like the male mallards for the same reason that female mallards do, hence the frequent hybridization. I note that the mallard x black hybrid in the photo is from Ottawa, Canada, where I live. A lot of us wonder how many pure black ducks are left here–female hybrids would be hard to detect.

      1. Is there evidence that more closely related species are more effectively hybridized? This would be an obvious guess, but I wonder if it was explicitly measured.

  2. Very interesting stuff. With birds, I often wonder if I’m looking at a hybrid or an immature specimen or a local variation or a seasonal change in plumage.

    But I wonder about similar studies involving sunfish. I’ve fished some ponds in which the catch is puzzling indeed.

    1. This undoubtedly has something to do with the daffy nature of ducks, along with their impulsive nature (as in the Disney mallard known as Donald).

    2. Well, it seems they may not be so slutty. At least, they have adaptations to reduce the likelihood of getting with egg (do they say that?). Google “female duck reproductive anatomy” and “corkscrew”. Don’t say I didn’t warn you.

  3. The “Canada Goose” in the photo with the goose hybrid is actually a Cackling Goose (Branta hutchinsii), a species that was split from Canada some years ago. I consider it a rather dubious species and wonder whether it will be relumped with Canada at some point, but so far it has maintained its full-species status.

  4. That Blue Heron / Egret hybrid was something else. Thanks for this interesting post. Hybrids occur a lot more often in captivity as breeders like to create novel “species”.

  5. I found this a very interesting read. Bird plumage can be confusing at the best of times, especially here where we are into an early spring.

  6. “Hybrids could be capable of having offspring but might have the wrong appearance or behavior to attract mates.”

    This sort of mating discrimination sounds like it would fascinating to investigate but hard to study in the wild. One feels sorry for the poor hybrids regarded as freaks by their almost-birds of a feather.

  7. I had worried that many hybrids ducks I’ve seen seem to be derived fro domestic species. I’m glad of the fact that species are shown to be very stable and largely incorruptible from hybridizing over time.

  8. I recently saw a very large and oddly proportioned egret in Malibu. It seemed so out of the ordinary both in size and (to a layman’s eye) unusual egret features that I stopped to take photos. I wonder if it was a hybrid.

  9. I’m sorry I can’t give a citation, but I recall a presentation at an ASIH meeting that reported that ~5% of offspring of redd-spawning fishes in streams in NW Pennsylvania, all tributaries of Lake Erie, are hybrids, some inter-generic.

    Birds copulate so females have a little control over which male sires their young. Except, perhaps, for birds like ducks whose males force copulations. These fishes, however, release sperm and eggs over/in gravel in streams. Small wonder that some sperm find the wrong eggs.

  10. It seems to me that hybridization must be rare and often the hybrids sterile practically by definition. If it were not the case, how could the inter-breeding species remain distinct?

  11. Indeed, a very interesting paper. We already knew that hybridization is relatively common on a species level. A few years ago, I estimated that about 16% of bird species has hybridized in the wild.

    However, estimating the incidence on an individual level is very difficult. This study provides a good try, but I think they might be underestimating it.

    Anyway, for those interested in bird hybrids, feel free to check my blog Avian Hybrds.

  12. This morning on the “Earth Sky News” is an article by Julian Avery, an ornithologist, on feeding wild birds and he touches on a study of some hundreds of species that found that 75% of species had “side partners” and that birds were exceptionally “promiscuous” and then mention that they were “cheating”.
    And your observation in the paragraph. called “The Upshot” you referred to the “Sluttiest” species.
    I object!
    Attaching our moral values to other species, for shame.
    They are doing what comes naturally, naturally.

  13. BTW, Ebird is a great site. If you download their app on your phone and register, you can enter your location and call up birds people have sighted nearby as well as entering your own sightings.

  14. I’m confused by the overall and corrected hybridization calculation. The formula for overall would be (unique+dups)/(Total+dups) I think. Since the corrected would be unique/total, I would expect that number to go down, not up. e.g. (4+1)/1001=.4995% and 4/1000=.4%. What am I missing?

  15. Another fascinating and excellent post on evolutionary biology, from a real expert on species. Clearly the notion of “species” is an idealization, one which continues in use because it is a useful approximation. This study shows that in birds it is a very good approximation. It would be important for someone to find two distinct species that have a low but measurable rate of hybridization, and do large-scale genomics on their populations. We would either see the two species gradually merging, or see an equilibrium between the introduction of alleles from the other species and their elimination by natural selection. (For all I know this may be being done, I don’t keep up with the relevant literature).

Leave a Reply