The Arbornaut, page 12
Finding an insect in the forest canopy seemed akin to locating a needle in a haystack, and I was discouraged. I had successfully found one important herbivore in the Antarctic beech canopy but knew that almost every other tree species had one (or more) leaf feeders. How could I design fieldwork to discover some of the most important marauders? Most munching was a cryptic event, with insects both camouflaged on the leaf surfaces and feeding alone, making them really hard to find. Based on ecological literature, I knew insects were actually incredibly abundant in vegetation, yet elusive. Maybe it was like birdwatching, where years of practice leads you to develop an eye for the quick flutter of wings. Instead of spying on feathered flyers, I needed to learn to spot small, scurrying feeders on leaf surfaces. “Biodiversity” is the collective term describing the variety of species on Earth, of which many are six-legged, aka arthropods. In the 1800s, Charles Darwin estimated that Earth housed eight hundred thousand species. (I can only imagine the queen of England gasped and was most impressed with what seemed an enormous number at that time, as calculated by her prodigal young naturalist!) Nearly one hundred years later, a scientist at the Smithsonian Institution named Terry Erwin increased Darwin’s original estimate almost thirtyfold, just based on beetles he counted living in a tropical tree. Erwin sprayed pesticide overhead using a handheld mister to bring the arthropods down to the forest floor, where he counted the insect “rain” that fell onto a plastic sheet spread out under a tree canopy. From his fogging episode, Erwin extrapolated that the world may contain approximately thirty million species, and the majority were insects that had not previously been classified, a healthy majority of which were beetles. And because over half of all insects are reputed to be herbivores, ecologists now estimate that over 50 percent of the world’s terrestrial biodiversity lives in the canopy with its gazillions of leaves. In 1988, Professor Edward O. Wilson, the distinguished entomologist at Harvard University, upped Erwin’s estimates and speculated that Earth contains over a hundred million species, including bacteria and soil organisms. Recognizing that scientists have not adequately explored either the treetops or the soil ecosystems, Wilson believes both regions contain many undiscovered species with the potential to raise his estimate further still. But the speed of discovery of new species in rain forests, as I soon discovered, is extremely slow due to that “needle in the haystack” phenomenon, so scientists estimate that approximately 90 percent remain undiscovered; hence the term “eighth continent,” coined independently by several of us arbornauts, reflects the undiscovered expanse of forest canopies. Many of those undiscovered species may disappear without humans ever knowing. Even more mind-boggling is the number of insects living on our planet, estimated at ten quintillion (written as 10 plus eighteen zeroes). Of this total, a huge proportion of those six-legged critters live in the eighth continent.
Once I started focusing on the millions of unknown insects in tree crowns, I also started to appreciate the multitude of insects in our daily lives. It turns out insects are abundant yet invisible in almost every ecosystem, not just the eighth continent. How many of us know the number of arthropods that inhabit a cubic yard of our own backyards? Or the amazing creepy-crawlies living in our own homes with us? And most definitely, we have no idea how many insects live in forests, except to admit it is a big number. Science has advanced with so many astounding calculations: the distance to the moon, the diameter of an atom, the dimensions of a dinosaur, and even a map of the human genome. But we have not yet accurately counted the number of insects in a tree crown. Exploration of the eighth continent lags behind coral reefs, deserts, polar regions, and even outer space because only a handful of professional arbornauts exist, so we must scramble to catch up.
Once I realized insect herbivores were a huge threat to foliage and their feeding activities were difficult to observe, I needed to figure out some new methods to up my game as an amateur entomologist. My monthly field trips to observe leaves and chronicle their fate expanded to include a bug component. I had to sleuth out their consumption: who, when, how much, and how often. I added vials and insect nets to my field gear. In addition to swinging in a harness to sample leaves and then insects, I developed a pretty regular field-trip routine: pack, drive, sleep, wake up, climb trees, measure, photograph, catch insects during daylight, climb again, measure, monitor, shower (cold only), dinner, chase resident quolls (Dasyurus viverrinus, a marsupial relative of the Tasmanian devil that eats both insects and small animals) from stealing my dinner off the barbecue, sample insects by dark, collapse into sleeping bag, and fervently hope no one stumbles into my remote campsite after a long, liquid night at one of the rural pubs.
During the first year of fieldwork, I learned the skills of a good detective spotting clues of herbivores, ranging from frass to feeding signatures on a leaf. During my second year, I became woefully aware that herbivory resulted in significantly less chlorophyll, reducing the tree’s capacity to photosynthesize. After solving the mystery of the Antarctic beech host-specific beetle defoliator, I puzzled over an ongoing episode of coachwood defoliation in the subtropical crowns at the Dorrigo National Park field site. Almost every new leaf suffered some level of damage, but I rarely observed even one insect in the act of chewing. Then, during a routine monthly sampling trip, serendipity struck. I always camped in a tiny expedition tent because no hotels existed nearby except for a few rough pubs. The camping area in Dorrigo was called Never Never, an apt description because there was never anyone else at this isolated spot. I usually had both the outhouse and lone picnic table to myself, shared occasionally with a few bowerbirds and curious parrots. Male bowerbirds built courtship structures from sticks on the forest floor and decorated them with blue fruits and flowers to woo a female. Sadly, many of the bowers at Dorrigo were cluttered with blue drinking straws, litter from fast-food outlets. I always felt scared when a car drove into the parking area, but my biggest fear was failing to find an insect herbivore or losing track of a marked branch along the vertical rope transect. One starry summer night in February, I awoke at about 2:00 a.m. to use the outhouse. My footsteps crackled on dry leaves underfoot as the only other sound in the forest. Pausing to appreciate true darkness in this remote place, I heard some raucous, grinding noises overhead that reminded me of a truck shifting gears. In the woods, that was downright scary. I returned to the tent for a flashlight and entertained visions from that favorite childhood fairy tale of Jack up his proverbial beanstalk. As I aimed a narrow beam of light into the greenery above, to my amazement, thousands of beetles were munching on coachwood leaves, their metallic carapaces reflecting in my flashlight beam. Eureka! The aerial salad bar was frequented by nocturnal insects. It made sense for most insects to feed at night and avoid predation by birds during the day. I had been searching during daylight hours, which explained why I was coming up empty-handed! This was an exciting discovery, and subsequently useful when future entomologists started to explore forest canopies in other parts of the world. Thanks to my bladder, I was one step closer to demystifying the complex interactions between insects and foliage.
From that day forward, I included night climbs as part of fieldwork. It was spooky to ascend alone in the dark, and I had to exercise utmost caution to minimize encounters with anything poisonous, toxic, or aggressive. There was no exact precaution for avoiding potential predators while climbing a tree in the dark, but careful advance surveillance with my headlamp alerted me to any eyeballs overhead (usually spiders), dangling “strings” of nasty wait-a-while vines (if you got hooked, they never let go due to backward-gripping spines), and occasional brown bumps, which were marsupial tree possums. What a special privilege to share a whole new world of nocturnal vegetarians, not only chrysomelid beetles as in the beech canopies, but also katydids, stick insects, caterpillars, and sucking weevils! The landscape of a leaf surface, aka the phylloplane, was ever changing. Insect feeding frenzies were in full force in the treetops, and I was one of the first to climb up there to observe them in situ!
Not all insects escape their enemies via darkness. Some use temporal strategies, undergoing synchronous hatching to devour leaves en masse during a short period of time before predators detect and then eat them all, or before a tree mounts its own defense by producing defensive chemicals. The Antarctic beech beetle followed this strategy. Beech composed 95 percent of the cool temperate canopy (hence termed monodominant), and one insect, Novocastria nothofagi, survived by engaging a feeding frenzy on young beech leaves followed by rapid metamorphosis. Its life cycle was a race against time: enemies such as birds or infections could not build up quickly enough to control the explosion of larvae in the victimized beech crowns. In contrast to beech beetles, other foliage feeders survived via a strategy of rarity. If you are a lone walking stick in a sassafras tree, predators probably won’t spot you. Or if you are a katydid feeding solo at night, you have a double insurance policy against discovery by your enemies: escape in space via isolation, plus the cloak of darkness.
Like most field biologists, I always worried about the accuracy of my field methods. One of the fears in designing research is the challenge of avoiding bias. Finding an insect allowed for no ambiguity: either it was counted or it didn’t exist. We simply call this “presence or absence.” Calculating defoliation requires multiple samples to create averages and thereby achieve objective data; it was especially challenging to obtain accurate estimates throughout the enormity of a three-dimensional canopy, as compared to the controlled conditions of a laboratory. It might be tempting to select leaves that were not eaten, reducing time and effort to calculate defoliation, but that would result in a biased sample. So how could I be sure that my leaf samples and herbivory calculations were an accurate reflection of the entire canopy? Ecologists often use a simple table of random numbers to generate unbiased selections of whatever needs subsampling. For example, if you want to sample ten leaves out of thirty, you need to number them all and then consult a random number table to generate your subsample selection. The purpose of subsampling is to save time and energy (you avoid measuring every single leaf in the forest) and to avoid bias such as picking the closest, prettiest, or least eaten leaves. I learned about the importance of avoiding bias from my statistics professor, who told us about an expensive laboratory trial to study the swimming speed of an important oceanic fish. Every month, the scientists picked thirty fish out of a large tank of three hundred individuals and measured how fast they swam in a special chamber. After two years and great expense, the experiment was canned. By simply plucking fish out of the tank, instead of numbering them and then selecting random numbers from a table, they had inadvertently chosen the slowest-swimming fishes—the ones easiest to catch. (A nightmare for those ichthyologists, but their story will stick with me for life!)
To accurately sample insect consumption in a tall tree, I needed to understand a whole lot about holes in leaves. What was the range of defoliation for different tree species? What age of leaf tissue was preferred by herbivores? Third, and perhaps most critical to all leaf detective work, how could I accurately calculate leaf damage in a whole tree, many made up of millions of leaves, without harvesting the leaves or measuring every single one? These types of sampling questions burned in my brain, and unfortunately no prior publications offered protocols. Another big leaf-worry literally kept me up at night: If a bug takes one tiny bite out of a young leaf, does the resulting hole expand when the leaf becomes full-grown? In other words, does 10 percent chewed on an emerging leaf remain 10 percent when it matures? To answer the “holey leaf” conundrum, I designed a simple field experiment to compare sizes of holes in young versus full-grown foliage. The equipment list was amazingly inexpensive: two-dollar paper punch, waterproof Magic Markers, tags for branches (to find them again each month), and a notebook. I found a treefall that created a small light gap where understory coachwood branches took full advantage of sun flecks. Due to the high light levels in the clearing, numerous new leaves flushed close to the ground instead of ninety feet overhead. I numbered nine to fifteen leaves on each of three branches of three trees, carefully including four age classes: young, mid-aged, mature, and senescent. I used the paper punch to clip 0.33 cm (exact area of one punch), 0.66 cm (two punches), or 1.0 cm (three punches) from random leaves, and placed the holes as beetles would chew, avoiding major veins, which insect mouthparts can’t easily bite.
I punched all the leaves and waited. Patience is a requirement for most ecological research. Some data collection takes years, even decades, to complete. In this case, it only took several months for all the foliage to reach maturity. I was excited to return and harvest the samples, all carefully labeled. Back in the botany laboratory, I used a digitizer (a gadget that measures the surface area of any two-dimensional object placed on a belt running over a laser beam) to calculate if the holes had grown, shrunk, or something in between. I remeasured each punched hole in the young leaves that had expanded, in the old leaves serving as controls, and in the mid-sized leaves that underwent moderate growth. For every sample, the holes grew proportionally larger as the surface expanded but remained the same percentage of total leaf area. This was good news—if 10 percent of a young leaf was eaten, it remained 10 percent as an adult. It was a relief to know I didn’t need to worry about sampling error when measuring herbivory on a young leaf versus mature; the proportional amount of damage remained consistent regardless of age. This also meant that expressing data as percentages was more accurate than calculating millimeters of holes. Ten square millimeters (0.4 square inch) of young leaf damage turned into forty square millimeters (1.6 square inches) when it quadrupled in size.
Another challenge that plagued me: How could I account for leaves that were 100 percent eaten? If an insect eats an entire leaf, there is no easy way to find and measure it. Fortunately, the dogged process of monthly observations over several years allowed me to determine exactly how many were entirely eaten. Insects usually left clues after eating a whole leaf, such as frass, silk, or a dangling petiole. If I had a leaf numbered 8 located between leaves 7 and 9, but then it was replaced by a pile of frass on the stem—voilà—number 8 had likely been eaten. Usually, petioles remained intact because they were too tough to bite. Typically, I found foliage half eaten in month one, three-quarters eaten in month two, and then completely devoured by month three. High winds also contributed to entire leaf removal, but it was easy to detect damaged foliage due to storms because the entire canopy was impacted. My long-term data set revealed that trees suffered three to four times higher defoliation than previously estimated by forest scientists who made quick surveys limited to the forest floor. Earlier forestry publications cited that forests incur 5 to 8 percent annual defoliation, but my results proved that trees sustain much higher damage. Canopies tolerate insect attack ranging from 15 to 25 percent annual leaf area losses, information that will aid conservation and forest management in modeling healthy forests and insect outbreaks. With the onset of climate change where insect outbreaks are predicted to be on the rise, it is important to gauge the threshold of tree resilience.
The material ingested by insects quickly recycles back into the soil via frass. Rain filtering through the canopy, called throughfall, washes the pellets to the forest floor where they are rapidly reabsorbed by the root hairs. This is an important pathway of nutrient cycling. Conversely, when uneaten leaves eventually fall to the ground, they undergo more gradual decomposition due to their large size and waxy surface. The foliage of four out of five of my research species required more than a year to decay on the forest floor. (Yes, I carefully measured their decay rates, by placing thirty mesh bags containing equal weights of leaf material on the forest floor, harvesting three each month for ten months to weigh them, and thereby tracking the decay rates!) Insect defoliation may be beneficial to a tree because the digested tissue falling to the ground as frass rapidly recycles the nutrients. Severe droughts may circumvent this presumed advantage of frass reabsorption on the forest floor if the soil dries out and surface root hairs die. But severe droughts and their associated heat waves also lead to more insect outbreaks and ultimately more frass. These pathways are indeed complex.
Unfortunately, unlike birds or fish, trees can’t shift their location when conditions become unfavorable. At some point, forests exceed a tipping point of stress from the trio of drought, warming, and insect outbreaks, and they die. In the past few decades, insect attacks have surged with hotter, drier conditions created by human-induced climate change. Throughout the 1980s, the topic of climate change was mostly limited to geoscience and climatological circles, rarely making its way to ecological professionals. It is hard to believe that, only thirty years ago, many disciplines lacked that perspective about global change. We were so consumed with figuring out how ecosystems functioned and how so many creatures coexisted that we did not connect the dots and interpret the climate warning signs. Yes, we missed the forest for the trees. Had tropical ecologists recognized the significance of global warming sooner, perhaps we would have conducted biodiversity surveys throughout different forest ecosystems, to establish baseline data before climatic extremes began jeopardizing species survival. Now scientists are scrambling to catch up, as many forests burn out of control or insect epidemics increase in severity and frequency. We cannot know how much is extinct if we never knew what lived in these forests in the first place.
In addition to the potential biases of sampling leaves throughout a vast canopy and the challenge of calculating herbivory of leaves totally eaten, a third dilemma of fieldwork involved human bias. What if some people see things differently than others due to poor vision or tendencies to exaggerate? One solution is to engage multiple samplers to collect data. Believe it or not, one person might be overly timid in sweeping insects into a net, underestimating the counts, thus creating bias—a big no-no. I learned how to minimize human error by engaging teams of volunteers, now known as citizen scientists, to wield nets and count insects. Over time, a few hefty, macho net swingers who broke every branch in their path offset the one or two timid introverts who captured almost nothing in their delicate swings. So, how did I coerce fifty people to enter the jungle and sample insects, you might wonder? Earthwatch! This innovative organization, headquartered in Boston, serves as a clearinghouse to match volunteers with scientific research expeditions. For my first-ever grant application, I applied for an Earthwatch grant, requesting volunteers to measure herbivory in Australian forests. Grants are the lifeblood of science; almost all research requires outside funding to pay for equipment, staff, travel, and even such esoteric things as snakebite kits or laboratory safety goggles. The funding for grants is highly competitive, and newcomers are usually considered less qualified than the old guard, making it tough to break into the system. Carefully answering all the questions, I submitted the proposal and waited anxiously. Several months later, I received a YES letter. Oh joy! Getting a grant is one of the biggest drivers of happiness (and success) for scientists. Getting a NO is one of the worst experiences, but it happens to everyone because most grant applications have only a 5 or 10 percent chance of success. After three decades of field research, I have been fortunate to obtain millions of dollars of grants, but I have been turned down for an almost equal amount.
