A serious miscalculation on Ecuadorean swamps

Submitted by editor on 27 March 2014.

Nigel Pitman

Collecting plant specimens in a swamp of Yasuní National Park with a Huaorani friend, 1998. Photo by Tomás Delinks.


The paper on Ecuadorean swamps that was just published Early View in Ecography this week is the result of a serious miscalculation.

To understand why requires going back to the late 1990s, when I was carrying out forest inventories in Amazonian Ecuador. My goal was to sample tree communities in the three major forest types of Yasuní National Park, but there was no way of knowing how many plots I needed in each. In the blithe way of graduate students everywhere, I plucked some numbers out of the air: 15 ha in upland forests, 5 ha in floodplain forests, and 5 ha in swamps. Each hectare took between a week and a month to sample, including fieldwork (measuring trees, collecting specimens), herbarium work in Quito (drying and identifying the specimens), and the 10-hour bus rides back and forth between the two. It took a couple of years to finish all the plots, and by August 1999 I was done.

I sat down to write my thesis – and immediately realized my mistake. There were so many interesting things to say about the 15 ha of upland forest data that I could have written two dissertations about them alone. As it turned out, the floodplain and swamp data didn't even make it into the thesis I turned in nine months later. In other words, I had set up 10 ha of flooded forest plots – done six months of fieldwork – measured and identified some 6000 trees – gotten a lot of field assistants' feet wet – for nothing.

A swamp tree plot in Yasuní National Park, 1998. Photo by Nigel Pitman.


It felt awkward. Over the next few years I kept trying to get to those data, but work kept pulling me in other directions: new fieldwork, new datasets, new regions of South America – even new swamp plots. For years I gave up hope of working on the Yasuní flooded forest data, but I never forgot about them. Part of the reason was that I had a moral obligation to do something useful with them, seeing that most of the work had been paid for by taxpayers. And in the brief time I had spent with the data we had glimpsed some fascinating patterns. But what kept my thoughts returning to swamps, even as I got farther and farther away from them, were the memories of the fieldwork.

A Huaorani schoolboy who tagged along with our team in the swamps, Yasuní National Park, 1998. Photo by Tomás Delinks.


Really they were the most pleasant of the three forest types to work in. Walking down from the upland forest and into a swamp felt like crossing over into a cooler season – partly because the understory was spacious and open, with a breeze that rippled the palm fronds around you, and partly because you were up to your knees in water. The whiny cloud of mosquitoes that followed you everywhere in the upland forests was gone too – without them it felt like you could stand up straighter and see things more clearly. And while at first it was hard to wade through the murky water and not think about giant watersnakes, we never saw any. Once, in a swamp plot in Yasuní, I almost grabbed a little yellow and black snake coiled up at the base of a palm because I thought it was a roll of string I had dropped. When I realized my mistake and pulled my hand back, a shiver ran from head to boots and my friend gave me that significant look they give the bomb expert who almost cut the wrong wire. But that was the only snake I remember seeing in the swamps.

A few years ago, when I did finally find the time to look at the swamp data, two developments made the analyses easier. First, co-author Juan Guevara had methodically re-checked and updated all of the species identifications in the Yasuní dataset, and it was cleaner than it had ever been. Second, I had just read and was actively pondering over Robert Ricklefs' 2008 paper 'Disintegration of the ecological community.' That paper, which you should probably read instead of ours, contained one line in particular that I kept turning over in my head: "Indeed, the presence and relative abundance of a species at a particular point might depend on interactions with populations that do not occur there".

But as we lined up the data using the rough blueprint provided by Ricklefs (2008), we repeatedly found ourselves running into dead ends. It was as though we were trying to build a house but lacked the right tools. This beam joined that one, but where was the wrench that fit that peculiar bolt? In fact, an early, discarded draft of the paper ended like this:

More plant collections, Yasuní National Park, 1998. Photo by Tomás Delinks.

We close with some thoughts on what we learned in answering Ricklefs’ (2008) call to ‘disintegrate’ local communities by examining how the populations of their component species behave at larger scales on the landscape. Although the exercise yielded valuable insights, we were struck by the frequency with which answering the relatively simple questions posed in this paper required establishing arbitrary thresholds, carrying out non-standard analyses, inventing new terms, finding novel ways to visually represent information, and otherwise struggling with a lack of analytical tools (e.g., the absence of a standard statistical method to determine species preferences across more than two forest types; see Chazdon et al. [2011]). While our experience was positive, it also suggested that despite decades of advances concerning the role of spatial scale in plant ecology, the field remains undeveloped for explorations of the intertwined patterns of species abundance, frequency, and distribution across heterogeneous landscapes.


The good news is that the last few years have seen a burst of research on swamps of western Amazonia, including great papers by Outi Lähteenoja, Ethan Householder, John Janovec, Euridice Honorio, Katy Roucoux, Michael Gilmore, Bryan Endress, Luisa Ricaurte, and others, most of which you can find in the reference list of our paper.

On behalf of all the co-authors, and with thanks to everyone else who contributed in some way, I hope you enjoy the paper.



Chazdon, R. L. et al. 2011. A novel statistical method for classifying habitat generalists and specialists. – Ecology 92: 1332–1343.

Ricklefs, R. E. 2008. Disintegration of the ecological community. – Am. Nat. 172: 741–750.