CJ Battey







I'm a grad student in the Klicka Lab at the University of Washington Department of Biology and the Burke Museum of Natural History. The goal of the lab is to archive and analyze avian diversity. Sometimes that means catching birds and measuring stuff. Usually it involves extracting DNA and seeing how genomes vary between species and individuals.

My dissertation work is focused on using population genomic analyses to study how species respond to environmental change, both on historic (hundreds of years) and geologic (tens to hundreds of thousands of years) timescales. I'm currently working on measuring genetic variability in populations of hummingbirds that have experienced large range shifts in the last hundred years, hoping to better understand how the process of colonization affects genetic diversity in novel habitats. This winter I received funding from the NSF DDIG (!!) to conduct a comparative demographics study across a clade of hummingbirds (the Bees); using coalescent models to assess the timing and magnitude of changes in population size over the last c. 200,000 years for species distributed along a latitudinal gradient. I also work more generally on species delimitation and phylogenetics in birds, and occasionally venture into ecological data analysis of citizen-science projects. Producing beautiful and informative data visualizations of natural history observations is a major near-scientific hobby that I pursue probably more than I should.

Click around up top to read more about my research or see some pictures of my past fieldwork.


Population Genomics of Hummingbird Range Shifts: Diversity and Gene Flow at an Expanding Range Edge

In 1924 the famous California naturalist and biogeographer Joseph Grinnell described the Anna's Hummingbird (Calypte anna) as a bird of the Baja Peninsula and southern chaparral, breeding as far north as San Francisco with a few individuals dispersing up to near the Oregon border by the end of summer. Today Anna's Hummingbirds are a common urban bird in Seattle and Vancouver (BC), with regular post-breeding dispersal as far as southern Alaska. Last Spring UW's ornithology class found a female incubating chicks next to Lake Union in early March, and as I write this in mid-January I can see two male Anna's Hummingbirds fighting over the feeder in my neighbor's yard. The species has gone from vanishingly rare to ubiquitous in developed areas around Puget Sound in just 50 years, with most of the increase in the last two decades.

Over the same period Seattle's other common hummingbird (the Rufous Hummingbird, Selasphorus rufus, has also experienced a major range shift. Historically wintering in central Mexico, the species now occurs regularly throughout the winter on the US Gulf Coast - a shift of roughly a thousand miles. These range shifts probably have a number of causes - invasive and ornamental plantings, climate change, and hummingbird feeders, to name a few - but the process is likely to become increasingly common among many species in the coming century. The range shifts seen in the Anna's and Rufous Hummingbirds are some of the most dramatic observed in any species, and are relatively well-documented in survey records. This abundance of data allows us to study the population-level dynamics of a colonization and expansion process that many species are likely to undergo in the coming years.

In order to better understand the dynamics and impacts of range shifts on genetic diversity in bird populations, I am collecting tissue samples of Anna's and Rufous Hummingbirds from both core and novel ranges and applying a statistical modeling technique called approximate Bayesian computation (ABC) to compare different demographic models for how species colonize new areas. Do long-distance dispersal events and subsequent founder effects restrict the diversity of novel habitats? Or does gene flow with historic range cores cause edge populations to reflect the standing diversity of the whole range? How do heritable migratory orientations impact range expansions? Where did these range-shifted populations come from in the first place? My study should help us to better understand some of the basic natural history of these species, and can help us develop models for how we expect populations to behave as human impacts on the earth's landscape and climate increase over the next century.

This project recently received NSF funding for expanded sampling of the Anna's Hummingbird, and I expect to have preliminary data in a few months. Stay tuned!

The Genetics of Timing and Orientation in Migratory Hummingbird

The Rufous Hummingbird (Selasphorus rufus) makes the longest migratory journey of any hummingbird, with some individuals flying a circuit of over 7,000 miles between central Mexico and southern Alaska. The species is near-ubiquitous in riparian forests across the Northwest during breeding season, but during their southward fall migration they are concentrated in the two western mountain ranges that provide the most abundant wildflower blooms for quick refueling: the Sierra Nevada and the Rockies. We know from previous (excellent) work by Dr. Lynn Carpenter that Rufous Hummingbirds roughly double their body weight during migratory stopovers in the Sierra Nevada - implying that they lose half their body weight during the first half of their migration. With that kind of physiologic stress, these birds probably have very little wiggle room in finding a good spot to refuel on their way south. So, if you're a fledgeling hummingbird born in Alaska about to head south on a 3500 mile solo journey (both parents leave the nest site before chicks start migration), how do you know where to go?

In some species of songbird, captive breeding experiments have shown that migratory orientation is heritable, and it seems to be passed on in a bizarrely straightforward way. If you breed a Eurasian Black-cap that migrates southeast with one that migrates southwest, their offspring will try to migrate straight south (Helbig 1991). Splits in migratory orientation like this have been found to be associated with population divergence and putative sympatric speciation in some species (e.g. Eurasian Black-cap Sylvia atricapilla, Swainson's Thrush Catharus ustulatus), but no one knows whether the same pattern occurs in hummingbirds, and our understanding of the particular mechanisms of heritable migratory preference is still in its infancy. My project aims to test for genetic divergence in Rufous Hummingbirds migrating along different routes, to associate migratory routes with breeding areas, and to use sequencing data from across the breeding range to describe the biogeography and demographic history of the species.

Right now I'm preparing DNA sequencing libraries from hummingbird tissues from natural history museums around the country, and working on optimizing DNA extractions from feathers to expand sampling in regions where populations are declining. This winter I'll be generating and analyzing sequence data to try to map migratory connectivity in the species (which breeding ranges uses which routes) and testing for significant divergences in genes associated with migratory orientation and timing in other groups of birds. In the meantime I've been working up a few different ways of using citizen-science data to visualize the cool looped migratory routes of the species (see the gif in the top right and maps below - both depict annual changes in report frequency for S. rufus in eBird data).

*data via eBird: eBird Basic Dataset. Version: EBD_relFeb-2014. Cornell Lab of Ornithology, Ithaca, New York. February 2014.

Migration and Gene Flow in the Eye-Stripe Vireos

Vireos are one of the last families of songbirds that don't yet have a good phylogeny (a tree describing the relationships among all the species and their common ancestors), so the Klicka Lab has been working for while to put one together. Our paper describing relationships across the whole family using mitochondrial and Z-linked DNA was recently published in Molecular Phylogenetics and Evolution (pdf here). I'm currently focusing my research on a small subset of the tree where it looks like we have two previously unrecognized species that have been incorrectly lumped with other birds with which they do not appear to exchange genes at all - one basic metric of what makes a species a coherent evolutionary entity. The group I'm looking at is the red-eyed vireo clade - a set of five species spread across the Americas. Here's a map of the breeding and wintering distributions for 4 of the species (the last is restricted to a small island in the Atlantic off Brazil):

The three widespread species are at least partially migratory, and they spend the winter in the same general area of northern South America. In summer, the black-whiskered vireo flies north and breeds across the Caribbean and in the mangroves spreading up the Florida coast. The yellow-green vireo heads back up to the mountains of Mexico and Central America. The red-eyed vireo is the really odd one - some populations stay in Colombia and Ecuador all year round, some fly south and breed in northern Argentina and parts of Bolivia (this is called an Austral migration), and some gear up for a long trip across the gulf of Mexico and up through North America as far west as Seattle and as far north as the Yukon.

Looking at the range maps and life history of these species, it seems weird that this single species of red-eyed vireo can use all these different strategies and areas. For a songbird that weighs as much as a handful of paper clips, flying from Colombia to Alberta is a serious physiological stress. We might expect that evolution would have led to some differences in birds that have this major life-history checkpoint (the migratory red-eyes) relative to those that don't (the sedentary red-eyes), and that birds breeding on different continents at different times of year would either show signs of highly restricted gene flow or represent the very recent development of a novel migratory behavior. Our analysis of mitochondrial DNA gave us the first indication that there was something wrong with the current taxonomy of the group:

I color-coded the species here, but the take-home is that the current scheme for naming species in the group is missing a lot of existing diversity. Both the yellow-green vireo (V. flavoviridis) and the red-eyed vireo (V. olivaceous) are split into two lineages (East vs West Mexico, and North vs. South America, respectively). In the Red-Eyed Vireo, all the North American birds are long-distance migrants, while the South American birds are a mix of short- to medium-distance migrants and sedentary (nonmigratory) birds. So is it a signal of divergence along migratory divides? Random chance? An artefact of the piece of DNA we looked at? Here and in general the answer requires more data.

Last summer I sequencing tissue samples from these populations using a newish hot-shit sequencing technology called ddRADseq. Suffice it to say that you get a ton more data with a much wider distribution across the genome than we get with techniques that target single genes. I'm still analyzing this data and writing up a paper for publication, but I have a few early results. To start, here's the actual tree of relationships in the red-eyed vireo clade, based on the vast majority of genetic evidence:

What this analysis tells us is that the South American red-eyed vireos shared a common ancestor with the black-whiskered vireo more recently than they did with the North American red-eyed vireo - in scientific terms, the red-eyed vireo is paraphyletic with respect to the black-whiskered vireo. The Northern and Southern populations currently recognized as the red-eyed vireo also do not appear to exchange genes at all - bayesian clustering, coalescent modeling, and nonparametric D tests all fail to find any significant introgression after the point of lineage divergence. This is very strong evidence that these two groups of birds are independent evolutionary lineages and should be recognized as different species. It also demonstrates how direct analysis of gene flow can improve genomic methods of species delimitation by evaluating criteria with a clear biological relevance that is sometimes difficult to discern in existing methods based on estimating how well data fits a multispecies coalescent model in which species split instantly with no gene flow after divergence (ie BPP, BFD, BFD*).

Range Expansion and Reticulate Evolution in the Great-Tailed/Boat-Tailed Grackle

The great-tailed (Quiscalus mexicanus) and boat-tailed (Quiscalus major) grackles are large noisy blackbirds native to Mexico/Central America and the northern Gulf Coast/Florida, respectively. If you've ever been in a K-Mart parking lot in Houston and felt weirdly nervous about the swarms of croaking glossy birds that look a little like small crows - those were grackles. Great-tails are common anywhere humans live in their range. Boat-tails tend to be restricted to the general vicinity of coastal marshes. They look the same, except that boat-tails have yellow eyes and great-tails have dark eyes (weirdly, the difference in eye color is only consistent in the areas where the two species coexist).

Over the course of the last century both species have expanded their ranges northwards into the United States, and great-tailed grackles are now among the most common birds in developed areas from Texas to southern California. I'm looking at the genetics of these closely related species (actually "this", but we'll get there) as they expand northwards because the recent range expansions are a rare opportunity to observe secondary contact between long-isolated lineages - one of the key stages in the process of speciation.

DNA sequenced from grackle feathers suggests that the great-tailed and boat-tailed grackles form a paraphyletic species complex rather than two good species - that is, that some great-tailed grackles are actually more closely related to the boat-tailed grackle than to other great-tails. Specifically a molecular phylogeny built with the Bayesian MCMC program BEAST analyzing sequences from the ND2 mitochondrial gene of grackles from across their range shows that the population of great-tailed grackles native to western Mexico (from Sinaloa north) diverged from the clade including both all other great-tailed grackles and the boat-tailed grackles about 900,000 years ago. Here's a simplified phylogenetic tree:

Each circle is a bird we sequenced, color-coded by clade (lineage, basically). Over in the southwestern US up into Nevada and parts of southern California you can see where the two clades are mixing, and apparently interbreeding freely. The weird thing is that boat-tailed grackles and great-tailed grackles almost never interbreed (almost). So we have an old divergence resulting in what appears to be little ecological differentiation and no apparent reproductive isolation between the two great-tailed clades; and a relatively recent divergence leading to lots of reproductive isolation and both ecological and morphological divergence between the eastern great-tails and the boat-tailed grackle. Pretty weird.

There are a lot of questions to explore in the system, and I'm working in a few directions. First, I'm using niche modeling to examine the relative amounts of ecological divergence between lineages. Second, I'm conducting some classic population-genetics and newer ancenstral-area-reconstruction analyses to take a more detailed look at the historical exchange of genes among grackles in Mexico, which is where most of the action seems to be in the deep divergences of the group. Last, I'm hoping to get a study together to see if the two clades of great-tailed grackles with a new zone of secondary contact in the western US are actually interbreeding freely. More to come.


A random subset of recent pictures. Find more on my tumblr


Download PDF

I also wrote this website from scratch with html, css, and a little jquery.


R scripts for dealing with RADseq data analysis, mapping occurrence data, climatic niche modeling, etc. are up on my github



CJ Battey
PhD Candidate, Klicka Lab
University of Washington Dept. of Biology
548 Kincaid Hall, Box 351800
Seattle, WA 98195-1800