The SOP that opens with the research question: a sample PhD statement of purpose review

I want to understand why some plant populations at the trailing edge of their range keep going for decades after the climate has moved past them.¹ Warming should be erasing low-elevation populations of montane plants, and the models say it is, but when you walk the sites the populations are often still there: smaller, older, producing fewer seedlings, and not dead². My working hypothesis is that established plants condition their own soil, through the microbial communities they recruit, in ways that let their seedlings survive conditions that seed from anywhere else cannot. I want to test whether that conditioning is real, how long it lasts once the adults are gone, and whether it can be moved. Those are the questions I would bring to your program.³

I did not start with that question. I started with a transplant experiment that mostly died. In my final undergraduate year I ran a reciprocal transplant of silvery lupine (Lupinus argenteus) across three elevations in the Wasatch Range, six hundred seedlings I had grown from wild-collected seed. The design was standard and the question was standard: is the low-elevation population locally adapted? In the second week of May a late frost hit the lowest site and killed sixty-one percent of the transplants there in one night. I had a spreadsheet with a lot of zeros and a thesis due in six weeks.

What I did was go back and census every individual, dead or alive, and record where it sat relative to the adult lupines already in the plot. I had not planned to collect that variable; I collected it because it was the only new information I could get with a tape measure.⁴ Survival after the frost was not random. Seedlings within about forty centimeters of an established adult survived at roughly twice the rate of seedlings in open ground, after accounting for aspect and canopy cover. My thesis ended up being a survival analysis of a frost I had not wanted, and the effect I found was one I did not know how to explain. Microclimate was the obvious candidate. Soil was the one I could not rule out, and it nagged at me.

After graduating I spent two years as a field and lab technician at a long-term research station in the Sierra foothills, on a project studying oak regeneration. The project had a greenhouse, and my supervisor let me use a corner of it for the question I had brought with me.⁵ I set up a soil inoculation experiment: lupine seedlings in sterilized soil, inoculated with live soil from under adult lupines, from open ground at the same site, or from a high-elevation site, and grown under two watering regimes. I ran it for fourteen weeks.

The first run was useless, and it was my fault.⁶ I had placed the pots in shared flood trays, and when I sequenced the sterile controls at week ten, they had a microbial community that looked a lot like the live-inoculum treatments. The water had moved the microbes. I had known that was possible in the abstract and had convinced myself it would be too slow to matter.⁷ I threw out fourteen weeks of data and rebuilt the experiment with individually bottom-watered pots in sealed saucers, and I added sequenced controls at weeks two, six and ten so that the next contamination, if there was one, would show up before the end rather than after. The second run held. Under the drought regime, seedlings inoculated with soil from beneath adult lupines survived at sixty-eight percent, against forty-one percent for open-ground inoculum and thirty-nine percent for sterile soil. Under normal watering there was no difference. The effect exists, at least in pots, and it only appears when the seedlings are stressed, which is exactly the condition a trailing-edge population lives in.

I know what that experiment does not show. It does not show mechanism: I do not know whether the effect is fungal, bacterial, or something about soil structure that I did not measure. It does not show persistence: I do not know whether the conditioning outlasts the adult that made it. And a greenhouse is not a hillside. Those three gaps are the shape of the dissertation I want to do. The first needs amplicon sequencing paired with targeted exclusion treatments (fungicide, filtered inoculum). The second needs a chronosequence of sites where adults have died at known times, which I have started mapping from the station's long-term plot records. The third needs a field inoculation experiment at real trailing-edge sites, which is the hardest and the one I care about most.⁸

Your department is the right place for this because of how two people there work, not because of where it ranks. Dr. Ingrid Sallow's lab runs reciprocal transplants across elevation with soil manipulations nested inside them, which is the design I want to build and have so far only built in a greenhouse. I have read her group's papers on plant-soil feedback at range margins closely enough to know where I would want to push: her transplants have used bulk soil, and I want to separate the microbial fraction from the physical one.⁹ Dr. Rafael Ossining's group models microbial community assembly with hierarchical occupancy models that handle exactly the detection problem I ran into when my sequenced controls came back messy. I can run a field experiment and I can sequence a plate. I cannot yet build the statistical model that tells me which of the four hundred taxa on the plate matter, and his lab is where I would learn to. I have not spoken with either of them yet, and I would rather they judge the question on the page than in an email.¹⁰

I have also spent two seasons doing the unglamorous parts of long-term ecology: rebuilding a fence line, re-flagging plots after a burn¹¹, and entering thirty years of oak seedling counts from paper datasheets into a database that finally lets people ask questions of them. I am not romantic about fieldwork. I like it because it is where the models get checked.

My goal after the PhD is a research position, in a university or an agency, where I can keep asking whether populations that should be gone are gone¹², and what is holding them there. I am applying with a question I can state in one sentence, two experiments that partly answered it and partly broke, and a clear idea of what I would do next. I would like to do it in your department.