Revisiting Harrison et al. 1988

In a paper published in The American Naturalist in 1988, Susan Harrison, Dennis Murphy and Paul Ehrlich demonstrate the existence of a metapopulation of the Bay checkerspot butterfly, a rare species whose larvae feed on plants that grow on patches of serpentine soil. Using a combination of field surveys, simulation modelling and historical information, Harrison and colleagues showed that the distribution of local populations is best predicted by features of the habitat and distance from a source of colonists. Twenty-nine years after the paper was published, I asked Susan Harrison about her the origins of her interest in this topic, her memories of fieldwork, and what we have learnt since about metapopulations of the Bay checkerspot butterfly.

Citation: Harrison, S., Murphy, D. D., & Ehrlich, P. R. (1988). Distribution of the bay checkerspot butterfly, Euphydryas editha bayensis: evidence for a metapopulation model. The American Naturalist, 132(3), 360-382.

Date of interview: Questions sent by email on 8th December 2017; responses received by email on 22nd December 2017.


Hari Sridhar: Correct me if I’m wrong, but this paper seems to have come out of work you did during your PhD at Stanford. Could you place this piece of work in the larger context of your PhD topic? How did you get interested in the Bay checkerspot butterfly and the topic of metapopulations?

Susan Harrison: While a field technician and then a Masters student at UC Davis, I became interested in three loosely-related subject areas. One was plant-insect interactions, which my advisor Rick Karban studied, and from which I learned the excitement of doing simple field experiments to test how species affect one another. Another was habitat fragmentation, which was a hot theoretical topic at the time, and which my honorary co-advisor Jim Quinn was studying. The third was the ecology of serpentine soils, which are fascinating on a pure natural history level because of the unusual flora and fauna they support.

My PhD advisor, Paul Ehrlich, and his associates had long studied the population ecology of the Bay Checkerspot butterfly, a relatively rare species whose larval host plants grow on serpentine soils. When I visited Stanford as a prospective doctoral student and described my combination of newfound interests, Paul and his lab manager Dennis Murphy unrolled a map of southern Santa Clara County showing many scattered outcrops of serpentine soil, some of which were known to support populations of the butterfly, and most of which had not been explored for it yet. I remember looking at that map and instantly recognizing it as the study system for me. Although it was over a year later when I first heard the term metapopulation, I knew that local extinction and colonization of populations on habitat patches were theoretically exciting topics that were understudied in the field.


HS: This paper has three authors. Could you tell us how this group came together and what each person brought to this paper?

SH: Paul Ehrlich had studied the Bay checkerspot butterfly since around 1960 at the Jasper Ridge Biological Preserve on the Stanford campus (a locality that was not part of my study).   His students and post-docs continued studying various aspects of its evolution and ecology for several decades, again mostly at Jasper Ridge. Dennis Murphy completed his PhD in that lab in the early 1980s and then became the lab manager as Paul had become more involved in international conservation.  Paul and Dennis made my study possible in two ways: first, their tremendous knowledge of the butterfly’s basic ecology, without which it would have been impossible to interpret its metapopulation dynamics; and second, through their material support, such as field technicians.


HS:  This study was conducted “in the vicinity of Morgan Hill, California”. How did you choose this site? Do you continue to work in this site till today? When was the last time you visited this site? In what ways has this site changed from the time you worked there for this study? What’s the status of the Bay checkerspot butterfly in this site today?

SH: My study region was by necessity the region where nearly all the known populations, except the one at Jasper Ridge, were found. The butterfly became federally protected in about 1988, leading to funding and political efforts that have since saved many of the serpentine sites in the study region from development.  However, this protection also meant that research on the butterfly became constrained by both the need for permits and the heightened concerns of landowners. For those reasons, and also because commuting to the Bay Area for fieldwork would have been unpleasant, I decided, after beginning my faculty job at UC Davis in 1991, not to keep working in that region.  I have occasionally visited some of those sites, in my studies of serpentine plant communities.


HS: You discuss the role of the severe drought of 1975-77 in causing the extinction of many E. editha populations. Have there been any more severe droughts since then that have had a major impact on butterfly populations? 

SH: The drought of 2012-2014 was even worse than the one in 1975-1977.  I have heard that some of the butterfly populations, notably the very large at Coyote Ridge, survived and are still doing fine.  I don’t know about all of the small populations.


HS: If you to think back to the fieldwork you conducted during this study, what are your most striking memories? Do you remember how you marked butterflies then?

SH: It was a deep hunter-gatherer thrill to get up early in the morning, drive and hike to some of the relatively wild natural habitats on the fringes of the southern Bay Area, and explore them looking for butterflies. Bracing myself to ask landowners permission to work on their land was also a memorably scary but usually positive experience.  We marked butterflies with Sharpie pen stripes on their wings.

Susan Harrison catching butterflies in the field (© David Harrison)

HS: In a couple of places in the manuscript you cite unpublished work by C. Boggs and ‘Murphy & Weiss’. Could you tell us who these people were and if this unpublished work was published later?

SH: Carol Boggs was a PhD research scientist who studied butterfly nectaring, oviposition, and energetics.  Stu Weiss was a lab technician, who did most of the mark-recapture work for our paper, and who later earned his PhD in the Ehrlich lab studying the effects of climate on the butterfly. I believe they later published the work cited in my paper.


HS: You acknowledge a number of people at the end of your paper. Could you tell us a little more about how you knew them and their contribution to this paper?


  1. S. Weiss: See above. We keep in touch. He knows more about the ecology and conservation of Bay Area serpentine grasslands than anyone else I know these days.
  2. K. Holl: My first field assistant, and now a very successful professor of restoration ecology at UC Santa Cruz
  3. P. Foley, J. Quinn, S. Schoenig, K. Thorarinsson: Academic friends from my Masters days in Davis, who helped me with the statistics and modeling in the paper.
  4. J. Baughman: A contemporary grad student in Ehrlich’s lab, who helped by making comments on the manuscript. I heard he left science and became a police detective.
  5. J. Diamond: A famous author and ecologist who was one of our reviewers. In his signed review, he said the paper was better than he expected it to be from its pedestrian-sounding title, which was originally “Metapopulation dynamics of the bay checkerspot butterfly.” We adopted the alternative title he suggested.
  6. M. Singer: A former Ehrlich student, who became the leading expert on the evolution of host-plant preference in the checkerspot butterfly, that is, E. editha, the species that includes the subspecies bayensis, but also many other subspecies found across California. He recently retired from a long career at U. Texas. He gave me some helpful advice on butterfly ecology, though I can’t remember what it was.


HS: How long did the writing of this paper take? When and where did you do most of the writing?

SH: It took the whole summer of 1987, between my first and second years in grad school.  I wrote it in my office at Stanford; people didn’t have home computers back then!   Periodically, I’d print it off on a dot-matrix printer, tape some figures into it, and give photocopies to my coauthors and others for their comments.


HS: Did this paper have a relatively smooth ride through peer-review? Was The American Naturalist the first place this was submitted to?

SH: Yes and yes!


HS:  What kind of attention did this paper receive when it was published?

SH: Most notably, and to my infinite excitement and professional benefit, it resulted in an invitation to speak at the first-ever conference on metapopulation dynamics, held by Ilkka Hanski of the University of Helsinki, Finland.  In turn, this conference resulted in my writing a review paper on local extinction in metapopulations, which is still my best-cited paper (1991, Biological Journal of the Linnean Society).

HS: What kind of impact did this paper have on your career and the future course of your research?

SH: Combined with a follow-up paper (1989, Ecology) in which I examined dispersal and colonization in the butterfly more closely, a metapopulation modeling paper (1989, Oikos), and with the review paper mentioned above, this paper made me, almost instantly, an expert on metapopulation dynamics.  This was simply because there then existed a tremendous amount of theory on metapopulations, great interest in applying the theory to conservation, and few field studies.  So it led to my receiving interviews and job offers a short time after finishing my PhD.

Oddly enough, though, it was ultimately not a lasting research direction for several reasons.  One was that, as mentioned above, continuing to work on this butterfly was logistically not feasible.  And in my own further explorations, I was unable to find study systems that were a good fit to classic metapopulation theory.  So I moved on to other research subjects, both related and unrelated to spatial ecology.


HS: Today, 29 years after it was published, would you say that the main conclusion still holds true, more-or-less: “suitable habitat patches for the bay checkerspot butterfly remain unoccupied because they are too distant from a source of colonists. This constitutes evidence for the existence of a metapopulation with a “mainland and islands” or “Boorman-Levitt” structure”.

SH: Yes, I’ve seen no evidence to the contrary.


HS: If you were to redo this study today, what would you do differently?

SH: Write it a little more concisely!   When I recently reread that paper, the long-winded style reminded me of the enthusiastic but inexperienced person I was then.


HS: Looking back at Tables 4 & 6 of this paper, how well do you think the predictions of your study have been borne out?

SH: See next answer.


HS: In the paper you ask “What is closer to the truth, the catastrophic 50-yr extinction model, which predicts 7 more colonizations over the next three decades, or the continuous-extinction model, which predicts constant turnover of populations and an equilibrium number of 8 to 11 populations?” Today, nearly three decades after this paper, which do you think is closer to the truth?

SH: To be honest, we don’t know any more today than we did back then about the actual patterns of turnover in those small populations, or others like them.   To find out would require surveying all those habitat patches annually for a decade or two.


HS:  This paper has been cited over 400 times. At the time of the study, did you anticipate at all that it would have this impact? Would you know what is it mostly cited for?

SH: Metapopulation theory reached its height of popularity in the 1990s, when it seemed like the perfect tool for understanding how endangered species do or don’t persist in fragmented habitats.  For around 10 years after the paper came out, I received many speaking invitations, review requests, et cetera, that reflected the general enthusiasm for metapopulation dynamics. As many loose ends as there were in my study, it was still one of very few field metapopulation studies, because it is so hard to get the necessary data – things like how many habitat patches and populations there are in a large geographic area, and how often these populations blink in and out over a period of many generations. So I think that many citations to my paper occurred when people were studying something about a species in a patchy habitat, would refer to their system as a metapopulation, and would use my study as an example.


HS:  Have you ever read this paper after it was published? If yes, in what context?

SH: Sometime in the past year I reread it just for fun, as one way to kill time and avoid reading the national news.


HS: Would you count this paper as a favourite, among all the papers you have written?

SH: Certainly I would count the study as a whole as a favorite. Graduate school is such a formative time when everything you do counts indelibly toward your perception of yourself as a scientist and toward your actual success.  The combination of emotional, intellectual, and physical energy that went into that study probably exceeds anything I’ve done before or since.  And I was very, very fortunate to have the opportunity to work on that study system, with all of the knowledge resources of the Ehrlich lab as a foundation.


HS: What would you say to a student who is about to read this paper today? What should he or she take away from this paper written 29 years ago? Would you add any caveats?

SH: When I think of how this study could have gone differently, one thing I would advise a student is to be well-versed in theory and statistics when you go out in the field, because that helps you see interesting patterns and collect the right data.  For example, I might have decided to only include serpentine patches that “looked like butterfly habitat”, and then my analyses would not have come out significant.  But I figured out that by collecting data on all serpentine patches in the entire region, good and poor-quality habitat, near and far from the source population, I could do elegant statistical and theoretical modeling to prove what was in retrospect a pretty obvious result.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s