Art Shapiro's Butterfly Site

Phenology has interested me for going on half a century. I began keeping phenological records of butterflies as a teenager in Philadelphia. As an undergraduate at the University of Pennsylvania I took a community ecology course from Jack McCormick, who was under contract to do an ecological study of the Tinicum wetlands (near Philadelphia International Airport). I had been doing an informal faunistic study of the place, purely for the fun of it, and had tons of data. A summary of this work was ultimately incorporated into Jack’s report. The study had a significant phenological component. Harry K. Clench, a butterfly taxonomist at the Carnegie Museum in Pittsburgh and one of the founding members of the Lepidopterists’ Society, published a phenological study of the butterflies of the Powdermill Nature Reserve in southwestern Pennsylvania, and sent me a copy. As I recall, we had already corresponded about the occurrence of unusual Hairstreaks (Lycaenidae) in southeastern Pennsylvania, a subject on which I had published field notes at a tender age. Thus began a correspondence on phenology which continued until Harry’s sudden death. Harry was “into” curve fitting. He had a sine function that worked pretty well for Powdermill, but not for Philadelphia. I was very leery of the approach: prediction was useful, but not nearly so useful as a method that cast light on the underlying mechanisms. The parameters in Harry’s equations were not obviously biologically meaningful.

My interest in phenology burgeoned at Penn, along with historical biogeography and edaphic endemism; all would stay with me for life. Exposure to the botanists Edgar T. Wherry and John M. Fogg, Jr. helped a lot. They had worked together on the flora of Pennsylvania for decades. When I wanted to find a rare butterfly, I would go to Wherry’s monk’s cell in the basement of Leidy Lab and tell him its host plant. Although already quite an old man, he would often recall localities in detail: “Take Bryan’s Mill Road 2.5 miles north, turn left on Old Bridge Road…”—but if not, he always had an annotated dot map within reach. And he knew my plant bloomed in early April…

When I went off to Cornell for graduate school I approached the upstate New York fauna in the same way. I had already formed the notion of a systematic study of butterfly phenology as a “community” phenomenon. This flowed out of my experiences with Jack McCormick and Robert MacArthur. MacArthur was my undergraduate adviser as well as being arguably the most influential ecologist of his generation. But let me begin with McCormick: as a plant ecologist (trained by Murray Buell), he had introduced me to the work of the Danish plant ecologist Christian Raunkiaer. Raunkiaer had classified plant species into “life forms,” which – he argued – represented adaptations to climate. He predicted that similar climates would select for similar “life form spectra” at the community level. This prediction foreshadowed the International Biological Programme focus on convergence at the community and ecosystem levels, decades later. I was inspired by the simplicity and elegance of the argument and began applying it to phenology. (Later I found that plant phenologists had had the same idea and began applying it to real communities around the time I was born. Ah, me.)

As for MacArthur, he was in the process of revolutionizing thinking about resource partitioning, niche relationships, and the factors responsible for “community structure.” In his view – elaborated from the ideas of his mentor, G. Evelyn Hutchinson – community structure was largely determined by interspecific competition for resources; species “fit together” in combinations that minimized such competition. Harry Clench believed that adult phenological patterns at Powdermill reflected both competition for, and the seasonal availability of, nectar resources. MacArthur’s student Bob Ricklefs argued that time itself should be treated as a resource. Putting all these threads together, I wanted to use butterfly phenology to make Raunkiaerian predictions and to test the reality of interspecific competition, the putative explanation of Clench’s phenomena and, presumably, mine.

But I was a graduate student and grad students, at least reasonably good ones, do not linger long enough in one place to gather long runs of data. Moreover, circumstances favored an extensive rather than an intensive approach to the New York fauna. As luck had it, we experienced some extreme short-term weather while I was in Ithaca – enough to stimulate the formation of hypotheses, however untestable in the grad-school context. If I wanted to look seriously at phenology, I needed reasonable stability in my life.

Meanwhile, I got significant impacts from Bill (W.L.,Jr.) Brown, who with E.O. Wilson had coined the term “character displacement” for the evolutionary process of competition-driven niche differentiation, and from Dick (Richard B.) Root and David Pimentel, both of whom worked on the organization of host-plant-centered insect “communities” in the 60s and were concerned with such things as species/genus ratios. When I published my monograph on the New York State butterfly fauna, I included some general remarks on phenology, but nothing more.

After finishing my doctorate I found a job teaching ecology and field biology in the Staten Island branch of the City University of New York, then called Richmond College. Anticipating a normal interval to tenure, and faced with a limited area and fauna, I decided to do a model phenology study there. But it was not to be: the opportunity arose to move to the University of California less than two years later, and I took it. My wife and I published a faunistic paper for Staten Island (we reworked the turn-of-the-century study by William T. Davis) with some phenological content.

Now, however, I was finally in phenological seventh Heaven: a Mediterranean climate. I knew enough multivariate statistics to know that if I wanted to partition phenological variance among climatic variables, a Mediterranean climate was the place to do it. I knew that Mediterranean climates are of relatively recent geologic origin. They are delicately poised – geographically and mechanistically – between the mid-latitude climates of traveling storms (bringing precipitation in all seasons) and subtropical and tropical climates with air-mass weather and wet/dry, not warm/cold seasonality. Mediterranean climates partake of the characteristics of both on a rather unpredictable basis, veering from one extreme to another.

I discovered that Sacramento weather records went back to the Gold Rush. Studying them, I decided that a 5-year program of intensive sampling at several sites would probably give me the data I wanted; that is, there should be enough weather variability in a 5-year period to allow statistics to do their job. Five years, not accidentally, coincided with my prospective probationary period to tenure. If I were fired in the end, at least I would have gotten my study out of my California experience!

I spent the 1972 field season looking for suitable sites and learning the California fauna – and flora – in detail. My previous experience going back to Tinicum suggested that a sampling interval of two weeks was a good compromise between the desirability of intensive coverage and the competing demands of other responsibilities. Most importantly, it would allow me to capture the year-to-year variability of butterfly phenology in adequate detail for analysis.

But now I was a lowly Assistant Professor, with all sorts of duties and no clout. (A few years later, my Department voted to minimize teaching for first-year hires to give them maximal freedom to establish their research programs. But that was too late for me.) My Department Chair actually called me in for a friendly bit of advice: my colleagues never saw me around the office or lab and were wondering what I was doing. I assured them I was learning how to do my research -–and could only do so in the field.

I soon had low-elevation sites at Suisun, North Sacramento and West Sacramento but was too busy to do them all every two weeks. I gave up NS and WS and restored them years later when I had the seniority to exert a lot of control over my own schedule.

Robert MacArthur died tragically young, of cancer. His students and associates put together a symposium in his memory and I was invited to contribute, even though I had been a mere undergraduate appendage to the MacArthur group. I used the occasion to publish an overview and prospectus of my phenology study (it appeared in the book “Ecology and Evolution of Communities”, edited by Martin Cody and Jared Diamond, 1975). If you read it, you will see that the plan has been fulfilled and the initial results, reported there, have been sustained in the long run. What I could not have visualized then was that the study would continue for more than 35 years!

I did get tenure, but not because of this study. No one gets tenure for long-term research, which does not bear fruit on the tenure-decision time scale. One gets tenure by publishing what are known in the trade as “quick and dirty” studies – lots of them, and not TOO dirty.

Meanwhile, the weather acted up beyond my wildest dreams.

From the Gold Rush to the late 1870s the regional climate had been remarkably unpredictable. Sacramento rainfall varied between 6” and 36.” It then quieted down. But now it went on a tear again, and as the extremes piled up the data were too good to give up on. The study continued: ten years, fifteen, twenty... Meanwhile I had much more control of my schedule; the number of sites and the number of visits went up to a plateau, with me spending upwards of 200 days a year afield monitoring my transect. (See Table 1.) And then we started to lose biodiversity; concern about global climate change began growing exponentially; and here I was sitting on one of the biggest and most-relevant data sets on the planet, but too busy to sit down and start analyzing. Finally my grad students, led by Matt Forister, sat me down and persuaded me that the time was ripe to try for National Science Foundation funding to get everything into an accessible on-line database, which would greatly facilitate its digestion.

So here we are.

TABLE I. Growth of the transect study from 1972 through 2009.