“The model cuckoos were actually stuffed real cuckoos – taxidermic mounts, which we got from the museum. We only had two. Of course, reviewers these days could rightly point out that there’s lots of pseudo replication, that we’re using the same model again and again. I think we could defend that by saying even getting one stuffed cuckoo was pretty amazing, because these are protected birds, and we were relying on specimens that the museum allowed us to have.”

“The next step was to test whether a slight infection with a common fish parasite, Ichtyophthirius multifilis, the “white spot” disease, reduced the health condition and the red intensity of the fish. We thought that was easy to do, because among wild caught fish you usually find an infected one and can breed the parasites taken from that individual. Bad luck, this time we could not find any infected fish. So Theo and I went to town and looked in all aquaria shops for an infected fish. Only in the last shop we saw an infected fish. However, when the owner tried to take the fish out of the tank for us, he said “I cannot sell you this fish because it has parasites”. It took us quite some time to convince him that we needed the fish because of its parasites.”

“The one difference is that instead of going online to read papers by people, I had to go to the actual physical library and look through actual physical journals to read their papers. But otherwise, it’s not that different. And then, you wrote to people, and again, instead of emailing them, you had to write to them. But then you wrote to them, and they answered.”

“The paper is obviously rather primitive by current standards when it comes to plant phylogeny, but remains of high standard when it comes to insect data. Interestingly, we used these insect data to make some rather daring extrapolations in order to estimate global insect numbers. Despite the fact that these numbers are of interest to many biologists and non-biologists, we have not progressed much with making these estimates more precise over the past 14 years.”

“The paper was much longer in print than is normal for QRB [Quarterly Review of Biology], which imposed page charges. I was a poor graduate student who could not pay them. My PhD committee and some other faculty generously agreed to support me, and it cost them several thousands of dollars of their research budgets to do so. Those contributing included Con Wehrhahn, Bill Neill, Don McPhail, and Buzz Holling. Tom Northcote, Peter Larkin, and Charlie Krebs may have chipped in as well.”

“The phylogenetic approaches have changed a lot. Traditional approaches were based on Sanger sequencing. This approach provides long, high-quality sequences but suffers from a number of limitations including the ability to sequence only a single locus per reaction. In contrast to Sanger sequencing, Next-Generation Sequencing technologies provide vastly larger quantities of data much faster and for far less money. As a result, next-gen sequencing approaches have rapidly taken over. So, yes, the availability of data is really shifting things tremendously. But to tell that particular story, we didn’t need any more data than we showed in the 2004 Science paper.”

“The programs used to calculate food web statistics was created to run on Apple computers that are obsolete today, and the code was never revised for more modern machines. So it is not used anymore.”

“The reason why I study barn swallows, and why many other people have joined this effort, is because they are extremely abundant, very easy to handle, very easy to observe and very easy to catch. This might not seem like important advantages to a non-biologist, but, believe me, there are very few organisms where you can catch individuals and follow them throughout their lives. There are many organisms, including Drosophila, where you can’t do this. They might be the wrong size, or if they are the right size, you never see them again once you have captured and tagged them. In summary, I think you can even call it laziness, to choose a species that is easy to study. But if you want to get robust results you have to pick a model system that is easy to work on.”

“The reasons given for rejection ranged all over the place. Some of them were ad hominem attacks; a lot of them were of the gist, “I can’t believe this result; it just is not possible.”The Nature review was, “If you accept this paper, I will never review for you again.” One was, “Has Hubbell lost his mind? His theory cannot be correct because it’s based on a false assumption.” Things like that. They were mostly incredulous and highly skeptical and completely unhelpful. I mean, like, nobody even wanted to consider the idea. It was pretty universally rejected, and it was very discouraging. With the wisdom of hindsight, I’m much more mellow about it now. And if you consider this a paradigm shift in a way, that’s exactly what you expect from true believers.”

“The result has proved repeatable. In the new experiments we have four replicates, and we show that it’s happening, but we also have a much better idea about why it’s happening, than we did then. The aspect of that paper that didn’t hold up was that it was written around the idea that it was differences in age-specific mortality that caused the life-history patterns that we saw. [...] we now know that that’s not the explanation. Indirect effects of predators and density regulation are playing a very important role.”

“The same year I was hired in the Zoology department in Texas, just by chance, they hired a post-doc from Capranica’s lab – Walt Wilczynski – in the Psychology department. We knew each other from before, though we weren’t close friends. But we both felt that looking at this question – variation in the neural basis of sexually-selected calls – would be very much worth doing.”

“The situation is quite different now. Two years ago, Bill Rice and I edited a collected volume on sexual conflict, The Genetics and Biology of Sexual Conflict, and I wrote a review chapter for that volume on the theoretical literature on sexual conflict. The conclusion that I was getting there, from a theoretical perspective, was that there are at least six different types of dynamics that can happen in models of sexual conflict. The Nature paper identified only two of them – continuously coevolutionary change and random drift along the line of equilibrium. Subsequent work produced a number of other things. And then, in terms of Bill’s idea about sexual conflict being “an engine of speciation”, I think I end my review paper from two years ago by saying that, yes, indeed, it’s an engine of speciation, but it works only under the right conditions. If conditions are not right, the engine will just stall. So, the story, of course, is much richer today.”

“The tradition in Ecology at the time was that PhD students developed their own projects. They also applied for their own small sources of funding, which meant that their projects had to be cheap. Dissertations could be on very different organisms, and on very different questions, from those of their advisors. (For example, my advisor, John Hendrickson, worked on sea turtles.) Hence, most PhD students published solo-authored papers. This does not mean to say that our advisors were not essential. My advisor cared very deeply about his students and the quality of their work. I have tried to maintain this tradition in my lab group. Some of my students have run independent projects as part of their dissertations. I am not a coauthor on these studies.”

“The trees were large. Most of them were Anacardium excelsum – a relative of cashew – and they tend to get very large hollows. Most of the hollows were big enough that I could go into the tree. One roost was in a silk cotton (Ceiba pentandra) tree that was so big that I could lie down completely inside the tree. Inside the hollow, yes. The diameter of the tree was about six feet inside, probably about 10 feet outside. So these were big trees. The bats would be at the top of the hollow, which, in most cases, was pretty high – probably twenty to thirty feet above where I was. So, I would actually lie on my back and observe them with binoculars. I tried to video record them with Super 8 mm film but that didn’t work very well. With a camera you are stuck in one location, but the bats move around a lot so I needed to move around too to keep track of them. We also had to wear respirators when making the observations. Histoplasmosis is a fungus that grows in bat guano, something I was very aware of and concerned about. So the students that helped me and I always wore these respirators. It was not easy. It was very tiring so I always had someone helping me. One person would call out observations and the other person would sit outside the tree and write them down. We also made little devices that would emit a click at set intervals of time, which we would listen for with earphones. I think for all the vampire work, I had set it to go off every 10 seconds. At every click we recorded whatever behaviour the bat was doing. I got the idea for this device from a primate field course I took as an undergraduate student at UC Davis. Davis is one of the eight places in the US that has a primate centre, where different species of monkeys or apes are kept in large outdoor enclosures. During that course we observed bonnet macaques in one such enclosure and recorded behaviours with the help of these clicking devices. You know, at the beginning I was not sure exactly what I was going to find. I just knew I needed to be systematic, be quantitative. I knew I wanted to study food sharing but I didn’t know how often I would see it, so it was done with some level of hope. The information from these focal animal samples was mainly used in a subsequent paper where I report on social grooming. The food-sharing was so infrequent that I recorded it whenever I saw it. It was so rare that if I had done it only on focal animals I would have had no data. In fact, in the first six months, I think I had seen it only a couple of times. At that point I was starting to think that the whole project was doomed. I think part of the issue was that – and this is not mentioned in any paper –in order to see the animals and identify them, I used coloured reflecting bands – bird bands, basically – on the wings of the bats. And to see the bands inside the tree I had to shine lights on them. Initially, the bats would always hide from the lights and so I couldn’t make any observations. In order to overcome that, I would, every single day, take a miner light into the tree and shine it on the bats continuously. These would last for 12 hours with rechargeable batteries, and the bats had nowhere to go, so in some time they got habituated to the lights. I could confirm this with a night vision scope which I could use with infra-red light. Infra-red light is invisible to bats. So when I compared their behaviour with the night vision scope and with the lights I couldn’t tell any difference. But it took months.”

“The University of Florida had excellent facilities for analyzing soil. It’s a public land-grant institution with a large Agricultural Science component. I did all of the soil analyses in the UF soil testing laboratory. They taught me how to do the analyses. Tom Sinclair, who was in Agronomy (and a colleague of Missy Holbrook) helped me with the bureaucracy. That simplified everything. It would have been would have been very difficult to set up all of those analytical procedures at Harvard. At the University of Florida, it was all set up; I just had to learn the protocols and run them on my samples.”

“The wild tobacco plant grows in habitats of the Great Basin Desert, of Utah, Nevada, and Colorado. And this particular site, Lytle Preserve, is owned by Brigham Young University. They had a little station to which you could go back every evening after work. And because this was already established, pretty much in the middle of the habitat where wild tobacco plants grow, this was a straightforward choice. You have to consider that wild tobacco plants grow only when the late succession vegetation, which is a sagebrush-dominated community of plants, is burnt down. Tobacco is one of the pioneer plants that comes after a wildfire destroys the original vegetation. And so, you’re dealing with a situation where you have to find a new population every year. Every time you go out for field research, you have to find that one burnt area where wild tobacco plants are growing. And so, you need some sort of field base from where you can go to all the places that potentially have wild tobaccos.”

“Their visit was nicely timed, fortunately, as heavy rain fell while they were there. Therefore we know the exact date when the drought ended.”

“Then we fitted some log-normal distributions and log-series distributions just using a conventional method of regular goodness-of-fit statistics. But now there are more powerful model selection statistics, which can be used, and would strengthen insights into how well different distributions fit. That said, I don’t think it would change our conclusions about the patterns that we saw.”

“There are some aspects of this paper, particularly the way I displayed the data in Figures 2 and 3, which I think I intentionally oversimplified nature in a way that has always left me uneasy. (Note that there are no error bars …)”

“There is a now-famous true story about how the size of the large plots was set at 50 hectares. I asked Robin how big the plot needed to be. I told him that I had mapped 13 hectares in Guanacaste and that we found 120 species in those 13 hectares. We knew from [Tomas] Croat’s flora that BCI [Barro Colorado Island] had roughly 450 species over one centimetre in stem diameter. So I said, “Robin, you know, we’re going to have to do a much bigger plot than 13 hectares, because we just don’t know if we’re going to get enough individuals to study, of each species, if we don’t do a large plot." And so I said, "I think it’s essential that at a minimum we do a plot at least twice as big as the Guanacaste plot, maybe 25 hectares." And Robin without any delay, shot back, “No, we’re going to do 50.” And that’s the entire scientific rationale for 50 ha plots in the global CTFS [Centre for Tropical Forest Science] network! We had no idea what the abundance of species would be in 50 hectares when we started, and as it turned out, there were lots of really rare tree species in the plot, too rare for us to analyze their species-level demography.”