Reported by Mike Kaspari, Kyle Harms, and Jennifer Powers

Meetings are a vital part of the process of science. We meet with colleagues old and new, exchange ideas, schmooze, and testify. At the same time, these conferences are expensive in time, money, and carbon (who doesn’t feel a twinge of guilt when boarding a plane for an overseas meeting, or settling into the seat of an over-air-conditioned conference room?). Still, there isn’t a technology yet that can capture the feel of a lively discussion and reproduce the many random discoveries that arise in the hallways of a scientific congress.

Which doesn’t mean that we can’t be tweaking them a bit. There’s always room for improvement.

So here in Honolulu at the Association for Tropical Biology and Conservation meetings, we held a symposium “Exploring Elemental Limitation of Tropical Biological Processes Across the Entire Periodic Table”. But we decided to model it on the philosophy of Ignite/Lightning talks. The basic structure: 14 ecologists from a range of backgrounds were asked to prepare a 3-slide, 3 minute presentation of an idea or result that intrigued them and was relevant to the topic. After Kyle laid out the ground rules, the audience of 150 or so listened to a variety of points of view: both defenses and critiques of the traditional NP focus, evocations of neglected elements like B and Na, uneasy reports of little to show for multi-year experiments conducted on a shoestring, and confident long lists of findings in a 15-year experiment. The talks may have varied in their ultimate fit to the topic, but they certainly front-loaded the many facets of nutrient limitation in the minds of the audience. Jennifer diplomatically kept everybody to his or her 3 minutes. 50 minutes elapsed quickly.

Kyle called for a brief break to stretch our legs and partake in cheesy snack chips that rhyme with Bonitos . This was blatant bribery of participants and audience alike, and was amazingly effective.

Then I (Mike) moderated an hour’s give and take. For Jennifer and Kyle, and certainly for me, it seemed to start a little slowly (in retrospect, what discussion among people in a lecture hall doesn’t)? Silvia Alvarez-Clare’s presentation prompted early stab at a discussion catalyst: “What determines how long you wait for a fertilization experiment to show a response? That is, when do you give up?” The immediate reply from an audience member was “After your 3 year NSF grant is up”. Turns out, there’s nothing like a jaded riposte about grant funding to break the ice and warm up the crowd. All too soon, the hour was over.

Here are some nuggets that emerged from the discussion.

1. It’s awfully obvious, but must be said, the answer to “What are the mechanics of nutrient limitation?” depends on the question.

Although there was a great diversity of interest in the audience (and among the organizers) the modal participant seemed to be an ecosystem ecologist with a big interest in the response of trees to plots that have been fertilized. This was best reflected by Joe Wright’s workmanlike summary of the results from the flagship fertilization plots on Gigante. However, when the synthesis question was posed “Well then, what have we learned about the role of nutrients on carbon sequestration in tropical forests?” (this for the benefit of global change modellers in the audience) you could hear the crickets. A definite uncomfortable silence. Which leads to…

2. In a tropical forest ecosystem, decomposition is dominated by organisms about 20 o.m. smaller than the trees that are scrubbing CO₂ out of the atmosphere. This has consequences for what responds to fertilization and when.

The number of bacteria in a few square meters of forest soil is on par with the number of trees in the Amazon basin. The generation times of the collembolans and oribatids in that soil can be measured in weeks and months. Not surprisingly, folks who work on microbial ecology and soil invertebrates—the brown food web—report rapid, repeatable responses to fertilization. We are talking hours, days, and weeks to see a bump in decomposition or oribatid densities. Moreover, over the course of months and years, these communities rearrange (and seem to continue to be rearranging) themselves. [And there is much more to do: imagine Lensky type experiments where chronosequences of soil communities are archived and revived over the course of a 10-year experiment]. So, large-scale fertilization plots continue to be enormously useful tools for studying the community and macro-ecology of soil bacteria, fungi, and invertebrates. Put another way, these plots have allowed us to study both direct effects of fertilization at the outset (i.e, a pulse experiment) and cascading indirect effects as the fertilizer application continues over time (i.e., a press experiment). In contrast, these same fertilization experiments, to the trees of a mature forest, even after 5 or 10 years, may reflect only the beginning of a pulse experiment. Which begs the question…

3. So how do tree ecologists maximize the utility of fertilization experiments?

A number of ideas arose. First, exploit the considerable forestry literature for clues as to what nutrient mixes are most effective and the time scales of responses. Second, focus on the short-term plasticity. For example, we know that N, P, and K additions are relatively quickly reflected in leaf tissue (though the consequences of that reallocation are still unclear). Third, identify tree species a priori that should benefit from N, P, or some other nutrient, and experimentally explore nutrient niches of co-occurring populations. The last, and most intriguing idea to come out of this part of our discussion was to look at reproductive consequences. For example, does all that extra phosphorus and nitrogen wind up in the endosperm of seeds? This would have obvious consequences for seedling growth rates and hence the demography of the forest. An animal physiologist in the audience (who was obviously slumming it) prompted this avenue of inquiry when she brought up the notion of how early exposure to nutrients may, through developmental switches, have long-term consequences for the phenotype. But, of course…

4. Fertilization plots are only one way to test hypotheses of nutrient limitation.

Chronosequences, geograpical gradients in biogeochemistry, and greenhouse experiments are all building the case for the role of nutrients in regulating the life history, community composition, and ecosystem fluxes in tropical forests. A mix of those data may ultimately be the most useful tool for studying the collective responses of trees to biogeochemistry. If so, the tree ecologists who started these plots were doing a mitzvah for those of us who study the small things. We owe them beer. And co-authorships.

5. Stoichiometry varies at every organizational scale, and reflects, potentially, the demand side of the nutrient regulation equation.

Greg Asner showed how both growth form and taxon membership leaves a clear signal in the multi-element stoichiometry. Yadvinder Malhi suggests the canopy drives a plant’s stoichiometric signature.In a similar vein, Emma Sayer stressed that the diversity of resource use at every scale should cause us to question even the concept of nutrient limitation at the ecosystem scale. The symposium identified one clear hole in our knowledge: datasets that simultaneously map biogeochemical availability and elemental use at the individual, population, and functional level.

6. It’s sometimes difficult to see the forest ecosystem for all the trees.

Given the modal audience member (see above), it was at times hard to conceive of a tropical forest as little more than large autotrophs embedded in a mineral matrix populated by critters whose job it was to turn detritus into plant food. But not always. Joe Yavitt made a case that nutrients can shape the way soil aggregates form. These aggregates, in turn, shape hydrology and represent the physical structure of the microbial world. Kyle Harms was a particularly strong advocate for exploring higher trophic levels and interactions based on the very cool work of Adriana Bravo. Bravo has shown that figs—packages of bat food that are bribes for the dispersal of fig seeds—often do not concentrate enough sodium, such that strict frugivorous bats need to seek sodium elsewhere. Kyle asked, how often is sodium used by plants, who don’t much need it metabolically, as a lure for the plant’s pollinator and disperser associates? Likewise, Nina Wurzburger and Sasha Reed showed why nitrogen fixation—with its dependence on Fe, Mo, and S (and lots and lots of carbon) is the poster process for multi-element limitation. One came away impressed with the huge opportunity to build nutrient limitation—from the structure of enzyme systems through the nutritional ecology of bacteria, fungi, and animals—into food webs and nutrient cycling.

7. There’s more to life than N and P.

A number of participants, including Rebecca Ostertag and Ben Turner, argued why N and P co-limitation will remain focal points of ecology. Erika Marín-Spiotta shared patterns of key enzyme activity for N and P-cycling that varied by precipitation-season and habitat. One point of the symposium, however, was to explore the ecological implications of the 25 elements required for life. Lots of neat stuff here. Katie Heineman made the case for a sharper focus on calcium (where up to 50% in a tree may be sequestered in bark); Kyle argued for sodium; Brian Steidinger made the case for boron (and, more generally, metals like boron and selenium that yield a hump-shaped relationship between concentration and performance). Jennifer Powers combined the results from decomposition studies in the lab and nutrient use studies in the field to make a strong case for multi-element limitation as pervasive across tree species.

7. And finally, the way ahead.

Two members of the organizing committee spent a good portion of their talks arguing for new, and very different, protocols to study multi-element limitation. Jennifer offered one protocol modeled on drug-screening: Step 1–start in the simplest mesocosms with the widest combination of nutrients. Step 2–Keep only the most promising nutrient combinations and test those nutrients in a more realistic mesocosm. Step 3–take the most promising results and test those in a longer duration field experiment. In that way, we should be able to winnow the 25 possible limiting elements, and their combinations/compounds, to something more manageable.

I went holistic, arguing that soil fertility (essentially, the first PCA in any analysis of soil biogeochemistry) can account for nearly 50% of the biogeochemical variation across tropical forest soils. This measure of fertility suggested the promise of “Kitchen Sink Experiments”: a fertilization treatment consisting of all possible nutrients at levels found in the richest soils. KSE’s (pronounced “Kissies” ™) test an alternate null hypothesis of nutrient limitation: not that no nutrient limits, say, decomposition, but that all nutrients, when increased together, produce the maximum result. Imagine a fertilizer based on the richest soils, the highest end of the PCA1 axis, applied from white sands to rich volcanic soils. Could be fun.

So, did it work?

As Kyle, Jennifer, and I debriefed each other, participants, and audience members, we came away feeling pretty good. Sure, there were some awkward silences, and certainly the moderator could use some training in fundamental social skills. But by and large, as the summary above attests, a fair bit of ground was covered in 2 hours. More to the point, the true gauge of success comes when comparing our Incite session to the alternative: eight 15-minute talks, each with from 0 to 5 minutes for questions, each with a rotating audience. Instead, we delivered a smorgasboard of ideas that the audience was able to select among; we identified knowledge gaps; had some good ideas (has anybody looked at how fertilizers influence seed nutrients and performance?) and offered some ways forward. And Kyle came through with the Doritos. Given the many costs of scientific conferences, we think this symposium model–lightening/incite talks followed by moderated discussion–offers more potential to generate new ideas in an efficient and fun way. It should be a basic feature of conferences in the future.