Peter Marler published the first map of San Francisco's white-crowned sparrow dialects in 1970, in Behaviour. The map was a simple thing. Three colors. Three dialects. The Presidio dialect to the north, the Berkeley dialect to the east, and the Lake Merced dialect to the south.
The boundaries between them were sharp, almost street-corner sharp. A sparrow on the north side of Geary sang one song. A sparrow on the south side of Geary, a few hundred metres away, sang a different one.
This was the moment, more than any other, that bird-song dialect became a serious research subject in North America.
The Nuttall's subspecies, Zonotrichia leucophrys nuttalli, is non-migratory. It lives year-round in the coastal scrub and city edges of central California. The males stay put. The young birds learn from the males they grow up near. This is why dialect boundaries can hold.
The boundaries have not held everywhere, but they have held more than one might expect. A 2014 paper by Elizabeth Derryberry's group revisited Marler's transects and found that some of the original dialect markers were still present at fifty years.
Pell Murphy is not Derryberry. He is, in his own words, a hobbyist citizen scientist who has been keeping eBird checklists in the Bay Area whenever work brings him to the West Coast. In April he was in San Francisco for a conference and gave himself three mornings.
On the first morning, the eighth of April, he walked Lobos Creek in the Presidio. He heard six white-crowned sparrows in the first hour. Three were close enough to record clearly with his phone.
The Presidio dialect, as documented since Marler, has a distinctive trilled middle phrase. Murphy's recordings, played back later against Macaulay Library samples from 2002 and 2018, were consistent. The trill is intact.
On the second morning, he walked the dunes at Lake Merced. The Lake Merced dialect historically has a clearer two-note whistle in place of the trill, and a slower terminal phrase. He heard four males. The two-note whistle was present in three of them.
The fourth was odd. It had elements of both dialects. Murphy is cautious about making too much of this. Hybrid songs occur at dialect boundaries, and the south end of Lake Merced is near the contact zone with the San Bruno coastal population.
On the third morning he walked the eastern edge of Golden Gate Park, looking for the Berkeley-type dialect that Marler had mapped along the bay frontage. He did not find a clean example. The habitat in that part of the city has shifted enough since 1970 that the population may have shifted with it.
Derryberry's group has documented some of this. Urban noise has driven changes in the structure of white-crowned sparrow song across the Bay Area. The fundamental frequency has risen in noisy territories. Phrases have shortened.
During the 2020 lockdowns, when traffic noise in the city dropped, recordings showed sparrows reverting toward older, more elaborate song forms. The 2021 paper that documented this got more press than most ornithological papers and deserved it.
Murphy was not in San Francisco for any of that. He was in Asheville. But he read the papers, and on his April mornings he was listening for what fifty-six years of dialect history has and has not produced.
What he heard, with caveats, is this. The Presidio dialect is durable. The Lake Merced dialect is durable but porous at the southern edge. The Golden Gate Park population has shifted enough that the Berkeley type, as Marler heard it, is hard to find in the original transect.
The work that needs doing is not Murphy's to do. It is being done, methodically, by Derryberry's lab and by graduate students who are repeating Marler's transects on a schedule.
What the casual visitor can do is sit on a bench in the Presidio at six in the morning in April, with a phone in their pocket, and hear the same dialect that Peter Marler heard in 1969.
This is unusual. Most things in a city the size of San Francisco do not last fifty-six years.
The Nuttall's white-crowned sparrow has lasted because it is non-migratory, because it learns from neighbours rather than parents, because the Presidio's coastal scrub has been more or less protected since 1972, and because the boundaries between its dialects fall along habitat edges that, in this city, have been remarkably stable.
Murphy's checklists from the three mornings are filed with eBird, location-named and time-stamped, where they will join the long, undramatic accumulation of what amateurs in the Bay Area have been measuring for two generations.
The research story of this species is the citizen-science story in microcosm. One man's maps in 1970, a long line of students after him, and the rest of us, with phones and bench seats, hearing what they heard.
