Introduction
Citizen science is a system for involving interested laypeople and amateurs in scientific research projects. Cornell University Ornithology Lab was an early pioneer in citizen science. One of its major modern endeavors was Project PigeonWatch, which launched in 1993 and ran until 2012.
The Cornell project was meant to document the diversity of feral pigeon plumage patterns worldwide. This diversity contrasts dramatically with the monotypic blue-bar patterns of their wild ancestor, the rock dove still found in portions of the Middle East, the Mediterranean borderlands, and the coastal cliffs of northwestern Europe. Rock doves (
Columba livia Gmelin 1789), like many other wild species of bird, are driven toward a single morph by stabilizing selection against departures from the norm: Unusual morphs are easily tracked by a raptor among a scattering flock of birds. Feral pigeons, however, exhibit quite a range of common morphs, with variations in color and variations in patterns overlying their base colors. These variations are persistent against stabilizing selection, and Cornell wanted to know why (LaBranche 1999). Are they maintained by divergent selection, perhaps local adaptation to complex local habitats? What is the rôle of sexual selection and assortative mating in all this? Are divergent oddities in coloration favored during mate selection enough to counter stabilizing natural selection?
Project PigeonWatch enlisted thousands of laypeople worldwide to collect data with a simple template and reference materials. Participants counted the size of flocks they observed and the numbers of birds belonging to each of seven common morphs.
Additionally, if they observed courtship attempts and matings, they noted which morph displayed male courtship behaviors and which morph was the target.
Over the course of the project, refereed publications detailed the process of citizen science (e.g., Bonney 2007), described potential benefits of such large scale projects in science education at various levels (Bonney et al. 2009), or focussed on assessment of student learning outcomes and affective engagement of the public with urban birds through informal education (e.g., Melber 2006; Cullen 1999). No scientific results were ever published, however (cf. Allen and Cooper 2006). I recently approached the Cornell Ornithology Lab about these “dark data,” and they confirmed that the “science” in this particular “citizen science” project never made it into publication (Bonter 2018). This is, actually, a common outcome in many citizen science projects (Burgess, et al. 2017), perhaps in part due to the logistics of managing the volumes of data generated (Newman et al. 2011).
Among participants in PigeonWatch were certain of my students in biogeography and California ecosystems courses at Cal State Long Beach. Those who did not show up for course field trips had to do an “alternative self-guided urban fauna field trip” to watch pigeons in widely scattered locations. My intention was to send the data forms to Cornell, but I procrastinated until, when I finally assembled the forms, I learned Cornell was no longer operating PigeonWatch.
Data and Methods
Unable to contribute my students' data to this now defunct project, I decided to keep on having my students collect them anyway and then analyze them on my own. I now have data on 6,475 pigeons observed in 360 sites and 610 courtships among them, collected by 84 of my students using the PigeonWatch protocol over eighteen years. All students were asked to record their observation sites' precise locations. This summer, I transferred all the students' paper reports into an OpenOffice Calc spreadsheet, manually geocoding records in Google Earth.
Converted into a spatial database, I calculated the percentages of birds at each site that fell into the seven basic morphs, generalized them further into four classes (wild-type blue-bar, melanic or dark blue, red, and albinic), and then classified the flock by the dominant morph of these four, if any, or else as diverse. I then calculated several standard measures of biodiversity, using numbers of the original seven morphs instead of numbers of species. I also coded all courtships into one of 49 types.
To visualize all this, I created a Google Earth KMZ file. Each site was assigned a placemark icon colored by morph class and graduated by size of the flock.
Seventy percent of the flocks contain a dominant morph class, one with at least half the birds. Thirty percent of the flocks have no one morph class rising to such dominance, and diversity is more common in larger flocks. Dominance varies over space but in no intuitively obvious pattern. What governs such spatial shifts in dominance? Might differential predation shape the direction of natural selection to adjust the birds' crypticity to different land surface backgrounds?
A graduate student, John Rowles, clipped out quarter mile radius buffers around each site in our department's collection of National Agricultural Imagery Program images for Los Angeles and Orange counties. He focussed on more recent semesters coïnciding with the images in our archive, yielding 105 site files.
I processed the clipped images in GIMP to extract R, G, B, and greyscale means, medians, and standard deviations.
I then ran simple linear regressions between the lightness or darkness of the land cover and the percentage of birds in one of the four generalized morph classes, with the expectation that predation selection should drive local populations to be more similar to local prevailing land cover colors. Results were pretty similar across all bands, so I'll use median blue readings to illustrate results.
Using the standard deviations of the various bands as a proxy for visual diversity in the landscape, I ran a simple linear regression between standard deviations and the flock diversity measures. The expectation was that surface diversity should be directly associated with flock diversity. Again, I'll use the blue standard deviation and, for flock diversity, Simpson's Inverse Index of Biodiversity.
Lastly, I focussed on sexual selection, cross-tabulating all of the courtships to explore the relative frequencies with which the different morphs initiate courtships or are the targets of courtship attempts in comparison with their availability in the regional pigeon population. These were processed with Z tests of the differences of proportions. Sexual selection of morphs should lead to proportions significantly different from regional proportions of available morphs. Then, I cross-tabulated whether courters and courtees were different from one another or similar to one another. If assortative meting is going on, males and their targets should disproportionately resemble one another and that might maintain morphic diversity. This was tested with Chi-square.
Results and Discussion
The relative frequency of the various morphs is not significantly associated with the darkness or lightness of their habitats,
not for red birds,
not for blue-bars, and
not for albinic birds --
except in the case of the melanic pigeons. As the lightness of their habitats increases, there is a significant increase of these darker morphs, which is exactly the opposite of what I expected! While significant, the effect size is pretty trivial: Adjusted R squared is only 0.03. The birds are apparently not experiencing balancing selection for crypticity, which suggests that predation, especially from raptors, is not a common cause of pigeon mortality.
As the visual diversity of their habitats increases, the morphic diversity of the flocks very significantly decreases, again counter-intuitively! The effect size is weak, though, with an adjusted R squared of 0.12. There clearly does not seem to be predation-enforced divergent selection for blending into a complexly patterned environment.
In terms of the relative amorous activities of the pigeon morphs, wild-type blue-bar males are very significantly friskier than the other male morphs, initiating courtships at a rate disproportionate to their representation in the regional population. Melanic males are slightly under-active with a prob-value of 0.08. Females of all morphs are courted at levels not significantly different than the levels of available morphs.
Looking specifically at which male morphs courted which female morphs, a pattern of assortative mating does become apparent. Blue-bar and melanic males disproportionately bestow their attentions on hens similar to themselves, while the rarer red and albinic males disproportionately seek out hens different from themselves.
Broken out by the four morphs of courters and courtees, all four basic morphs disproportionately favor birds of their own morphs, with varying levels of reduced attention paid to the other three morphs. The effect size, however, is modest, with Cramér's V at 0.21.
So, with the large data set acquired incidentally in my classes, it appears that pigeons' morphic diversity is not a response to predator-enforced balancing or divergent selection to blend into their habitats. It may be partly an artifact of some degree of sexual selection. On the one hand, the wild-type blue-bar male is the most active in courtships while all the female morphs seem likely to garner male attention in numbers reflecting their regional availability. Matched by who is courting whom, however, a weak pattern of assortative mating attempts appear among the males, who seem more inclined to go after females who look like them than otherwise. If the females share that predilection, the morphic diversity of feral pigeon flocks may reflect a kind of divergent selection through assortative sexual selection.
Conclusions and Recommendations
In previous studies of sexual selection, the discerning sex is the female. To the extent she can raise her offspring without male assistance, she is free to favor oddities in males and that can become a snowballing pattern that leads to such bizarre male ornamentation as the peacock's. Pigeons, however, are a pair-bonding species showing very little if any sexual dimorphism, so conventionally understood sexual selection seems improbable among them. The data presented here, however, suggest a closer look at assortative mating.
In the future, observation protocols might focus on differentiating the courtship behaviors seen among “cruising” pigeons from those found between a bonded pair. Are bonded pairs actually similar to one another or are they divergent? The nature of the pair bond depends not only on the male behavior prioritized by PigeonWatch's protocols but on whether the targeted hens respond to these displays and how their choices result in similar or divergent couples. There is some evidence that pigeons of different morphs have different physiologies and advantages or disadvantages in successful production of fledged squabs and that squab survival in difficult conditions is promoted by a morphically divergent parent couple (Jacquin et al. 2012). Greater field attention of a more complex nature is needed, then, to detect female choice in response to male courtship and the structure of the resulting pair bonds.
Another issue pertains to something that I think Cornell overlooked. Gene frequencies in any species are a product, not only of introduction by mutation and of natural selection and its sexual selection variant, but also of migration. The constant low levels of odd morphs in pigeon flocks may reflect their constant reïntroduction from domestic flocks. Every pigeon keeper loses birds. Most of these will die in short order but a few will survive, find mates, and reproduce, importing their quirky gene material. As a next stage, I am considering approaching the Los Angeles Pigeon Club to inquire about the locations of its members, the morphs present in their flocks, and their histories of lost birds. Mapping this information may well demonstrate the source of gene flow into feral flocks!
References
-
Allen, Paul E., and Cooper, Caren B. 2006. Citizen science as a tool for biodiversity monitoring. Original English version of La Ciencia Ciudadana como herramienta para el monitoreo de la biodiversidad. In Especies, espacios, y riesgos, pp. 17-32. Ed. Irene Pisanty and Margarita Caso. Mexico City: Instituto Nacional de Ecologia, Universidad Nacional Autónoma de México. Available at https://ecmmons.cornell.edu/xmlui/bitstream/handle/1813/39286/CitizenScience_Allen_Cooper_2006.pdf?sequence=2
-
Bonney, Rick. 2007. Citizen science at the Cornell Lab of Ornithology. In Exemplary Science In Informal Education Settings:Standards-Based Success Stories, pp. 213-229. Ed. Robert E. Yager and John H. Falk. Arlington, VA: NSTA Press.
-
Bonney, Rick.; Ballard, Heidi; Jordan, Rebecca; McCallie, Ellen; Phillips, Tina; Shirk, Jennifer; and Wilderman, Candie C. 2009. Public participation in scientific research: defining the field and assessing its potential for informal science education. A CAISE Inquiry Group Report. Online Submission. Available at https://files.eric.ed.gov/fulltext/ED519688.pdf.
-
Bonter, David N. 2018. Personal communication (14 October).
-
Burgess, H.K.; DeBey, L.B.; Froehlich, H.E.; Schmidt, N.; Theobald, E.J.; Ettinger, A.K..; HilleRisLambers, J.; Tewksbury, J.; and ParrishJ.K. 2017. The science of citizen science: Exploring barriers to use as a primary research tool. Biological Conservation 208: 113-120. doi: 10.1016/j.biocon.2016.05.014
-
Conrad, Cathy C., and Hilchey, Krista G. 2011. A review of citizen science and community-based environmental monitoring: Issues and opportunities. Environmental Monitoring Assessment 176: 273-291. doi: 10.1007/s10661-010-1582-5.
-
Cullen, Daniel P. 1999. Scientific Literacy and Project PigeonWatch: Evaluation of Citizen Science Programs. MPS Project, Cornell University.
-
Jacquin, L.; Récapet, C.; Bouche, P.; Leboucher, G.; and Gasparini, J. 2012. Melanin-based coloration reflects alternative strategies to cope with food limitation in pigeons. Behavioral Ecology 23, 4: 907-915. doi: 10.1093/beheco/ars055.
-
Krasny, Marianne E., and Bonney, Rick. 2005. Environmental education through citizen science and participatory action research. In Environmental Education and Advocacy: Changing Perspectives of Ecology and Education, pp. 292-320. Ed. Edward A. Johnson and Michael J. Mappin. Cambridge University Press.
-
LaBranche, Melinda S. 1999. Why study pigeons? Birdscope 13, 3: 3.
-
Melber, Leah M. 2006. Learning in unexpected places: Empowering Latino parents. Multicultural Education (summer): 36-40.
-
Newman, Greg; Graham, Jim; Crall, Alycia; and Laituri, Melinda. 2011. The art and science of multi-scale citizen science support. Ecological Informatics 6, 3: 217-227. doi: 10.1016/j.ecoinf.2011.03.002
-
Silvertown, Jonathan. 2009. A new dawn for citizen science. Trends in Ecology & Evolution 24, 9: 467-471. doi: 10.1016/j.tree.2009.03.017