Meandering meerkats

Dispersal – the movement of individuals to a new location – is a complex process that ecologists divide into three stages: emigration (leaving the group), transience through an unfamiliar landscape, and settlement in a suitable habitat. Dispersal is fraught with danger, as dispersers usually have a higher chance of starving, of getting eaten by predators, and may suffer a low reproductive rate.  So why move?

The problem is that there are major issues with not moving.  First, if nobody disperses, population densities could increase alarmingly, putting strains on resources and increasing the incidence of disease transmission.  Second, if nobody disperses, close relatives would tend to live near each other.  If these relatives mate, there would be a high probability of bad combinations of genes being expressed, leading to developmental abnormalities or high offspring mortality (geneticists call this inbreeding depression). In social species, such as meerkats, Suricata suricatta, the issues are even more complex, as dispersal could break up social groups that work well together to detect predators or find resources.  Nino Maag and his colleagues explored what factors influence meerkat dispersal decisions, their survival and reproduction, and how those factors affected overall population dynamics in the Kuruman River Reserve in South Africa.


A group of vigilant meerkats. Credit: Arpat Azgul

Meerkats live in groups of 2-50 individuals, with a dominant pair that monopolizes reproduction.  While pregnant, the dominant female usually evicts some subordinate females from the group; this coalition of evictees will either remain apart from the group (but within the confines of the territory) and eventually be allowed back in, or else emigrate to a new territory. By attaching radio collars to subordinate females, the researchers were able to follow emigrants to determine their fates.


Nino Maag collects data in the Kalahari Desert while a meerkat, wearing a radio collar, strolls by. Credit: Gabriele Cozzi.

How does population density affect emigration rates of evicted females?  You might think that meerkats would be most likely to emigrate at high population density, as a way of avoiding resource competition.  As it turns out the story is more complicated.  First, individual females (solid lines in graph below) are more likely to remain with the group (not emigrate) than are groups of two or more females (dashed lines). Second, emigration rates were highest at low population density, intermediate at high population density and lowest at intermediate population density. This nonlinear effect can be explained by low benefits of remaining in a very small group, so evictees are more likely to emigrate.  But as population density (and group size) increase, then the meerkats enjoy higher success as a result of cooperation between individuals  (in particular, detecting and avoiding predators).  But when population densities get too high, there are not enough resources to go around, and evictees are more likely to emigrate.


Proportion of evicted female meerkats that had not yet emigrated in relation to time since eviction at low (red), medium (light blue) and high (dark blue) population density.  Solid lines represent individual females, while dashed lines are coalitions of two or more females.

In addition to the density effects we just discussed, association with unrelated males from other groups early after eviction increased the probability that females would emigrate – presumably this increased the probability females would quickly create offspring in their new territory. Females also dispersed longer distances if unrelated males did not meet up with them, possibly to avoid inbreeding with closely-related males from neighboring groups.

Coalitions were more likely to return to the group if females were not pregnant – in fact 62% of pregnant evictees aborted their litters before being allowed back into the group.  Of the ones that did not abort before returning, only 42% of their litters survived to the first month.

The period of transience, when emigrators are seeking new territories can be prolonged and dangerous.  The mean dispersal distance was 2.24 km, and averaged about 46 days.  Larger coalitions with males present tended to disperse the shortest distances (left graph below). Dispersers took longest to settle at high population density – perhaps there were fewer available territories under those conditions (right graph below).


A. Effect of coalition size and presence of unrelated males on dispersal distance. B. Effect of population density on transience time (interval between emigration and settling).

Large coalitions settled more quickly than did small coalitions, particularly if accompanied by unrelated males.  Once settled, females successfully carried through 89% of their pregnancies (compare that to the 62% abortion rate of females that returned to their original group).  These females had a litter survival rate (to the first month) of 65%.

Social and non-social species are influenced by population density in different ways.  The situation is relatively simple for non-social species; as population size increases, competition between individuals increases, so dispersal is more likely.  However, even for non-social species, we might expect dispersal at very low population levels, if there are no mates available. For social species such as meerkats, the situation is more complex.  Cooperation enhances survival and reproduction, so it is better to be in a larger group (with more cooperators). At the same time, if the group is too large, then resource competition starts being an increasingly disruptive factor. As ecologists collect more dispersal data from other social species, they will be able to test the hypothesis that population density in many species influences dispersal in a non-linear way.

note: the paper that describes this research is from the journal Ecology. The reference is Maag, N. , Cozzi, G. , Clutton‐Brock, T. and Ozgul, A. (2018), Density‐dependent dispersal strategies in a cooperative breeder. Ecology, 99: 1932-1941. doi:10.1002/ecy.2433. Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2018 by the Ecological Society of America. All rights reserved.

Climate changes a bird’s life in shrinking grasslands

Back in graduate school, a couple of my grad student buddies and I would get together to fish for brown trout in the Kinnickinnic River in western Wisconsin.  We were students at the University of Minnesota (Twin Cities), but the Kinni was the closest trout stream.  Tired of catching small brown trout, we consulted a trout fishing map and discovered that the headwaters of the Kinni were rich in brook trout. So early one morning, map in hand, we followed strange paths and found our sacred brook trout haven. Alas, the only thing it was rich in was corn, now about two feet high – though there was a modest depression where trout waters once had flowed. Our personal depression was perhaps more than modest – having been robbed of brook trout, and the opportunity to experience some pristine waters flowing through a beautiful grassland.

Grasslands, one of the biomes native to parts of Wisconsin and Minnesota, are globally one of the most endangered biomes, because they usually are relatively easy to convert into farmland and suburban developments. Native grasslands harbor a wide biological diversity; consequently conservation biologists are concerned about their continued loss.


Cool-season grassland in southwest Wisconsin. Credit: John Dadisman.

Ben Zuckerberg, Christine Ribic and Lisa McCauley wanted to know how environmental factors influenced the nesting success of grassland birds, in particular, because as obligate ground nesters, they might be susceptible to changing  weather conditions that will be affecting the climate in coming decades.  A nest built on the ground is much less insulated from the environment than one built in or on a tree or even a ledge.

Bobolink 7 days (Carolyn Byers)

Seven day old bobolink chicks in a ground nest. Credit: Carolyn Byers.

Zuckerberg and his colleagues used Google Scholar and the ISI Web of Science to comb the literature (1982-2015) for studies that explored the nest success of obligate grassland birds in the United States. They identified 12 bird species from 81 individual studies of 21,000 nests. Based on their experience and the literature, both precipitation and temperature were likely to influence nest success, which is the proportion of nests that fledge at least one young. They considered three precipitation time periods: (1) Bioyear – previous July through April of the breeding season, (2) May of the breeding season, (3) June – August of the breeding season. They considered breeding season temperatures during May, and during the period from June-August. The researchers were also interested in the size of the grassland (grassland patch size), reasoning that a larger grassland might provide more diverse microclimates, so, for example, a bird might be able to find a dry microhabitat for nesting in a large grassland, even in a wet breeding season.


Map of the identity and location of species considered for this study.

The researchers discovered that both temperature and precipitation were important.  Nest success increased steadily with bioyear precipitation (Figure (a) below).  Presumably, more rain led to more plant growth and more insect survival, which would help feed the young.  Taller plants could also help shade or hide the nests. In contrast, nest success declined sharply with precipitation during spring and summer of the breeding season (Figure (b) and (c)). Heavy rains during the breeding season can flood nests, and also decrease the foraging efficiency of parents who might need to spend more time incubating nests during rainstorms. Lastly, extreme (low or high) May temperatures depressed nest success, which was highest at intermediate temperatures (Figure (d)). Egg viability depends on maintaining a constant temperature, and the parents may be more challenged to thermoregulate at extreme temperatures.  Temperatures later in the breeding season did not affect nest success.


Effects of (a) bioyear precipitation (previous July – April of the breeding season), (b) May precipitation during the breeding season, (c) June – August precipitation during the breeding season, and (d) May temperature on nest success. Shaded area represents 95% confidence interval.

But all is not straightforward in the grassland nest success world. These main findings about precipitation and temperature interacted with grassland size in interesting ways.  For example high bioyear precipitation, which overall increased nest success, only did so for smaller grassland patches (dashed line in top graph below), but not for larger patches (solid line).  Extreme May temperatures had different effects on nest success in relation to grassland patch size.  Low May temperatures were associated with high nest success in small patches (dashed line in bottom graph) and with low nest success in large patches (solid line).  High May temperatures were associated with high nest success in large patches, and with low nest success in small patches.


Predicted nest success of grassland birds in relation to bioyear precipitation (top graph) and May temperature (bottom graph) in relation to grassland patch size.  Solid lines represent large grasslands, while dashed lines represent small grasslands.  Shaded area is 95% confidence interval.

The researchers were surprised to discover that patch size affected how weather influenced grassland bird nesting success. Some of the patterns seem intuitively logical; for example, in unusually hot breeding seasons birds had higher nest success in larger grasslands than in smaller grasslands.  Presumably, birds were more likely to find a cooler microclimate for their nests in a large grassland.  However it is puzzling why in unusually cold breeding seasons birds had higher nest success in smaller grasslands. The researchers are planning a follow-up study to better document and measure the existence of microclimates in grasslands of different sizes, and explore how different microclimates influence the nesting success of vulnerable grassland birds.  Finding that warmer temperatures and drought generally reduce nest success to the greatest extent in small grassland patches is strong incentive for conservation mangers to establish large core grasslands as a tool to maintain bird populations in the wake of present and future changes to the climate.

note: the paper that describes this research is from the journal Conservation Biology. The reference is Zuckerberg, B. , Ribic, C. A. and McCauley, L. A. (2018), Effects of temperature and precipitation on grassland bird nesting success as mediated by patch size. Conservation Biology, 32: 872-882. doi:10.1111/cobi.13089. Thanks to the Society for Conservation Biology for allowing me to use figures from the paper. Copyright © 2018 by the Society for Conservation Biology. All rights reserved.


Recruiting rhinoceros

Despite their immense size and unfriendly disposition, we humans have done an excellent job decimating the black rhinoceros (Diceros bicornis) population from several hundred thousand individuals around 1900 to fewer than 3000 individuals by 1990. Three subspecies are extinct, one is on the brink, while even the most successful remaining subspecies, the south-central black rhinoceros, is extinct over much of its former range, or only found in nature reserves.  Humans prize its horns, which historically were used for making wine cups, ceremonial daggers, and a variety of medicines that purportedly revive comatose patients, detoxify infections, and aid male sexual stamina. Trade of rhinoceros horn has been illegal since 1977, but poaching abounds.


Rhinoceros horn products seized by the Hong Kong Government. Credit: U.S. Government Accountability Office from Washington, DC, United States [Public domain], via Wikimedia Commons

This population crisis has motivated conservation ecologists to evaluate the best approaches to conserving the black rhinoceros, and to restoring it to parts of its former range. Translocation, moving rhinoceroses from one location where they are relatively well-established, to another where they are extinct or at very low numbers, is a viable approach to restoring populations. However, translocation did not always work well, as some rhinoceroses died following translocation, or failed to reproduce even if they survived translocation. While a post-doctoral researcher at the Zoological Society in San Diego, California, Wayne Linklater recognized that there were large datasets collected by government and non-government agencies that might answer the question about what drives better rhinoceros survival and breeding.  Unfortunately some of these datasets were difficult to access or interpret, but Jay Gedir, Linklater and several other colleagues persevered and combed through the data to identify the factors associated with successful translocation.

Dales photos Oct04 006

Black rhino is airlifted to temporary captivity before being translocated to a release site. Credit: Andrew Stringer

Many translocation studies use a short-term measure of success, such as survival or fecundity (fertility) of the translocated individuals.  The researchers reasoned that the most important measure, when it is available, is the number of offspring produced by translocated females that survive to an age that they too can reproduce, which in the case of rhinoceros is four years. Based on existing long-term studies, Gedir, Linklater and their colleagues compiled the offspring recruitment rate (ORR), which combines the variables of survival, fecundity, and offspring survival to sexual maturity. They found that ORR was greatest when mature females were translocated into a population that had a female-biased sex ratio.


Offspring recruitment rate (per year) in relation to age at release and sex ratio bias of the recipient population. Female bias is greater than 60% female, male bias is greater than 60% male. Numbers above graph are sample size for each category. Error bars are 95% confidence intervals.

So why does sex-ratio matter?  The researchers are not certain, but females are subjected to considerable sexual harassment by very aggressive males, so a female-biased sex ratio may lead to less harassment and improved survival and reproductive success for translocated females.

Translocated juveniles took longer to produce their first calf after reaching sexual maturity than did adults after being released. Again this effect was stronger with a male-biased sex ratio.


Mean time (years) to produce their first calf after reaching sexual maturity for juveniles, or after release for adults.

Many females (47%) produced no surviving offspring. This pattern of recruitment failure was most common in juveniles, and least common in adults translocated into populations with a female-biased sex ratio.


Recruitment failure of translocated females in relation to age at release and sex ratio bias of the recipient population.

Several factors can cause recruitment failure: 23% of females died following translocation and 24% of surviving females never produced calves. However, calf survival to sexual maturity was a robust 89%, and this survival rate was independent of age at translocation or population sex ratio. Among surviving translocated females, juveniles were about twice as likely as young or old adults to fail to produce a calf, but were equally successful at raising the calves they were able to produce.

Linklater was surprised at how important sex ratios (and presumably social relationships) were, particularly given that black rhinoceroses spend much of their time alone or with one offspring.

Neto and Gwala size each other up

Neto (the ranger) and Gwala (the rhino) size each other up.  Credit: Wayne Linklater.

The study did not support a major role for many other factors that had previously been considered important in translocation success, such as habitat quality, population density, number of rhinoceroses released and reserve size. Based on their analysis, the researchers recommend that conservation biologists should translocate mature females into populations with female-biased sex ratios to reduce rates of recruitment failure.  If juveniles must be translocated, they too should be moved into populations that already have a female-biased sex ratio to reduce levels of sexual harassment by males after they mature.

note: the paper that describes this research is from the journal Conservation Biology. The reference is Gedir, J. V., Law, P. R., du Preez, P., & Linklater, W. L. (2018). Effects of age and sex ratios on offspring recruitment rates in translocated black rhinoceros. Conservation Biology32(3), 628-637. Thanks to the Society for Conservation Biology for allowing me to use figures from the paper. Copyright © 2018 by the Society for Conservation Biology. All rights reserved.

Homing in on the micro range

I’ve always been fascinated by geography. As a child, I memorized the heights of mountains, the populations of cities, and the areas encompassed by various states and countries. I can still recite from memory many of these numbers – at least based on the 1960 Rand McNally World Atlas. Part of my fondness for geography is no doubt based on my brain’s ability to recall numbers but very little else.

Most geographic ecologists are fond of numbers, exploring numerical questions such as how many organisms or species are there in a given area, or how large an area does a particular species occupy? They then look for factors that influence the distribution and abundance of species or groups of species. Given that biologists estimate there may be up to 100 million species, geographic ecologists have their work cut out for them.

As it turns out, most geographic ecologists have worked on plants, animals or fungi, while relatively few have worked on bacteria and archaeans (a very diverse group of microorganisms that is ancestral to eukaryotes).


Two petri plates with pigmented Actinobacteria. Credit: Mallory Choudoir.

Until recently, bacteria and archaeans were challenging subjects because they were so small and difficult to tell apart. But now, molecular/microbial biology techniques allow us to distinguish between closely related bacteria based on the sequence of bases (adenine, cytosine, guanine, and uracil) in their ribosomal RNA. Bacteria which are identical in more than 97% of their base sequence are described as being in the same phylotype, which is roughly analogous to being in the same species.

As a postdoctoral researcher working in Noah Fierer’s laboratory with several other researchers, Mallory Choudoir wanted to understand the geographic ecology of microorganisms. To do so, they and their collaborators collected dust samples from the trim above an exterior door at 1065 locations across the United States (USA).


Dr. Val McKenzie collects a dust sample from the top of a door sill. Credit: Dr. Noah Fierer.

The researchers sequenced the ribosomal RNA from each sample to determine the bacterial and archaeal diversity at each location. Overall they identified 74,134 gene sequence phyloypes in these samples – that took some work.

On average, each phylotype was found at 70 sites across the USA, but there was enormous variation. By mapping the phylotypes at each of the 1065 locations, the researchers were able to estimate the range size of each phylotyope. They discovered a highly skewed distribution of range sizes, with most phylotypes having relatively small ranges, while only a very few had large ranges (see the graph below). As it turns out, we observe this pattern when analyzing range sizes of plant and animal species as well.


Mean geographic range (Area of occupancy) for each phylotype in the study.  The y-axis (Density) indicates the probability that a given phylotype will occupy a range of a particular size (if you draw a straight line down from the peak to the x-axis, you will note that most phylotypes had an AOO of less than 3000 km2

Taxonomists use the term phylum (plural phyla) to indicate a broad grouping of similar organisms. Just to give you a feel for how broad a phylum is, humans and fish belong to the same phylum. Some microbial phyla had much larger geographic ranges than others. Interestingly, it was not always the case that the phylum with the greatest phylotype diversity had the largest range. For example, phylum Chrenarchaeota had the greatest median geographic range (see the graph below), but ranked only 19 (out of 50 phyla) in number of phylotypes (remember that a phylotype is kind of like a species in this study).


Box plots showing range size distribution for individual phyla. Middle black line within each box is the median value; box edges are the 25th and 75th percentile values (1st and 3rd quartiles).  Points are outlier phylotypes. Notice that the y-axis is logarithmic.

With this background, Choudoir and her colleagues were prepared to investigate whether there were any characteristics that might influence how large a range would be occupied by a particular phylotype. We could imagine, for example, that a phylotype able to withstand different types of environments would have a greater geographic range than a phylotype that was limited to living in thermal pools. Similarly, a phylotype that dispersed very effectively might have a greater geographic range than a poor disperser.

The researchers expected that aerobic microorganisms (that use oxygen for their metabolism) would have larger geographic ranges than nonaerobic microorganisms, which are actually poisoned by oxygen. The data below support this prediction quite nicely.


Geographic range size in relation to oxygen tolerance.  In this graph, and the graphs below, the points have been jittered to the right and left of their bar for ease of viewing (otherwise even more of the points would be on top of each other).

Some bacterial species form spores that protect them against unfavorable environmental conditions. The researchers expected that spore-forming bacteria would have larger geographic ranges than non-spore-forming bacteria.


Geographic range in relation to spore formation (left graph) and pigmentation (right graph).

Choudoir and her colleagues were surprised to discover exactly the opposite; the spore forming bacteria had, on average, slightly smaller geographic ranges. Choudoir and her colleagues also expected that phylotypes that are protected from harsh UV radiation by pigmentation would have larger geographic ranges than unpigmented phylotypes – this time the data confirmed their expectations.

The researchers identified several other factors associated with range size. For example, bacteria with more guanine and cytosine in their DNA or RNA tend to have larger geographic ranges. Some previous studies have shown that a higher proportion of guanine and cytosine is associated with greater thermal tolerance, which should translate to a greater geographic range. Choudoir and her colleagues also discovered that microorganisms with larger genomes (longer DNA or RNA sequences) also had larger ranges. They reason that larger genomes (thus more genes) should correspond to greater physiological versatility and the ability to survive variable environments.

This study opens up the door to further studies of microbial geographic ecology. Some patterns were expected, while others were surprising and beg for more research. Many of these microorganisms are important medically, ecologically or agriculturally, so there are very good reasons to figure out why they live where they do, and how they get from one place to another.

note: the paper that describes this research is from the journal Ecology. The reference is Choudoir, M. J., Barberán, A., Menninger, H. L., Dunn, R. R. and Fierer, N. (2018), Variation in range size and dispersal capabilities of microbial taxa. Ecology, 99: 322–334. doi:10.1002/ecy.2094. Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2017 by the Ecological Society of America. All rights reserved.

“Notes from Underground” – cicadas as living rain gauges

Given recent discussions between Donald Trump and Kim Jong-un about whose button is bigger, many of us with entomological leanings have revisited the question of what insects are most likely to dominate a post-nuclear world. Cicadas have a developmental life history that predisposes them to survival in the long term because some species in the eastern United States spend many subterranean years as juveniles (nymphs), feeding on the xylem sap within plants’ root systems. Magicicada nymphs live underground for 13 or 17 years, depending on the species, before digging out en masse, undergoing one final molt, and then going about the adult business of reproduction. This life history of spending many years underground followed by a mass emergence has not evolved to avoid nuclear holocausts while underground, but rather to synchronize emergence of billions of animals. Mass emergence causes predator satiation, an anti-predator adaptation in which predators are gastronomically overwhelmed by the number of prey items, so even if they eat only cicadas and nothing else, they still are able to consume only a small fraction of the cicada population.


Mass Magicicada emergence picturing recently-emerged winged adults, and the smaller lighter-colored exuviae (exoskeletons) that are shed during emergence. Credit: Arthur D. Guilani.

Less well-known are the protoperiodical cicadas (subfamily Tettigadinae) of the western United States that are abundant in some years, and may be entirely absent in others. Jeffrey Cole has studied cicada courtship songs for many years, and during his 2003 field season noted that localities that had previously been devoid of cicadas now (in 2003) hosted huge numbers of six or seven different species. He returned to those sites every year and high diversity and abundance reappeared in 2008 and 2014. This flexible periodicity contrasted with their eastern Magicicada cousins, and he wanted to know what stimulated mass emergence.



Protoperiodical cicadas studied by Chatfield-Taylor and Cole.  Okanagana cruentifera (top) and Clidophleps wrighti (bottom). Credit Jeffrey A. Cole.

Cole and his graduate student, Will Chatfield-Taylor, considered two hypotheses that might explain protoperiodicity in southern California (where they focused their efforts). The first hypothesis is that cicada emergence is triggered by heavy rains generated by El Niño Southern Oscillation (ENSO), a large-scale atmospheric system characterized by high sea temperature and low barometric pressure over the eastern Pacific Ocean. ENSO has a variable periodicity of 4.9 years, which roughly corresponds to the timing Cole observed while doing fieldwork. The second hypothesis recognized that nymphs must accumulate a set amount of xylem sap from their host plants to complete development. Sap availability depends on precipitation, and this accumulation takes several years in arid habitats. So while ENSO may hasten the process, the key to emergence is a threshold amount of precipitation over a several year timespan.

Working together, the researchers were able to identify seven protoperiodical species by downloading museum specimen data (including where and when each individual was collected) from two databases (iDigBio and SCAN). They also used data from several large museum collections, which gave them evidence of protoperiodical cicada emergences back to 1909. Based on these data, Chatfield-Taylor and Cole constructed a map of where these protoperiodical cicadas emerge.


Maps of five emergence localities discussed in this study.

The researchers tested the hypothesis that protoperiodical cicada emergences follow heavy rains triggered by ENSO by going through their dataset to see if there was a correlation between ENSO years and mass cicada emergences. Of 20 mass cicada emergences since 1918, only five coincided with ENSO events, which is approximately what would be expected with a random association between mass emergences and ENSO. Scratch hypothesis 1.

Let’s look at the second hypothesis. The researchers needed reliable precipitation data between years for which they had good evidence that there were mass emergences of their seven species. Using a statistical model, they discovered that 1181 mm was a threshold for mass emergences, and that three years was the minimum emergence interval regardless of precipitation. Only after 1181 mm of rain fell since the last mass emergence, summed over at least three years, would a new mass emergence be triggered.


Cumulative precipitation over seven time periods preceding cicada emergence.

The nice feature of this model is that it makes predictions about the future. For example, the last emergence occurred in the Devil’s punchbowl vicinity in 2014. Since then that area has averaged 182.2 mm of precipitation per year. If those drought conditions continue, the next mass emergence will occur in 2021 at that locality, which is longer than its historical average. Only time will tell. Hopefully Mr. Trump and Mr. Jong-un will be able to keep their fingers off of their respective buttons until then.

note: the paper that describes this research is from the journal Ecology. The reference is Chatfield-Taylor, W. and Cole, J. A. (2017), Living rain gauges: cumulative precipitation explains the emergence schedules of California protoperiodical cicadas. Ecology, 98: 2521–2527. doi:10.1002/ecy.1980. Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2017 by the Ecological Society of America. All rights reserved.


Prey populations: the only thing to fear is fear itself

In reference to the Great Depression, Franklin Delano Roosevelt is famously quoted as stating during his 1933 inaugural speech “the only thing we have to fear is fear itself.” Roosevelt was no biologist, but his words could equally apply to a different type of depression – the decline of animal populations that can be caused by fear.


Roosevelt’s inauguration in 1933. Credit: Architect of the Capitol.

Ecologists have long known that predators can depress prey populations by killing substantial numbers of their prey. But only in the past two decades or so have they realized that predators can, simply by their presence, cause prey populations to go into decline. There are many different ways this can happen, but, in general, a predation threat sensed by a prey organism can interfere with its feeding behavior, causing it to grow more slowly, or to starve to death. As one example, elk populations declined after wolves were introduced to Yellowstone National Park. There are many factors associated with this decline, but one factor is fear of predators causes elk to spend more time scanning and less time foraging. Also, elk tend to stay away from wolf hotspots, which are often places with good elk forage.

Liana Zanette recognized that ecologists had not considered whether predator presence can cause bird or mammal parents to reduce the amount of provisioning they provide to dependent offspring, thereby reducing offspring growth and survival, and slowing down population growth. For many years, she and her colleagues have studied the Song Sparrow, Melospiza melodia, on several small Gulf Islands in British Columbia, Canada. In an early study, she showed that playbacks of predator calls reduced parental provisioning by 26%, resulting in a 40% reduction in the estimated number of nestlings that fledged (left the nest). But, as she points out, Song Sparrow parents provision their offspring for many days after fledging; she wondered whether continued perception of a predation threat during this later time period further decreased offspring survival and ultimately population growth.

Song sparrow

The Song Sparrow, Melospiza melodia. Credit: Free Software Foundation.

Zanette’s student, Blair Dudeck, did much of the fieldwork for this study. The researchers captured nestlings six days after hatching , weighed and banded them, and fit them with tiny radio collars. They then recaptured and weighed the nestlings within a few hours of fledging (at about 12 days post-hatching) to assess nestling growth rates.


Banded sparrow nestling with radio antenna trailing from below its wing. Credit: Marek C. Allen.

Three days after the birds fledged, Dudeck radio-tracked them, and surrounded them with three speakers approximately 8 meters from where they perched. For one hour, each youngster listened to recordings of calls made by predators such as ravens or hawks, followed, after a brief rest period, by one hour of calls made by non-predators such as geese or woodpeckers (or vice-versa). During the playbacks, Dudeck observed the birds to record how often the parents visited and fed their offspring, and whether offspring behavior changed in association with predator calls. This included recording all of the offspring begging calls.


Blair Dudeck simultaneously uses a tracking device to locate Song Sparrows and a recorder mounted to his head to record their begging calls. Credit: Marek C. Allen.

Fear had a major impact on parental behavior. Parents reduced food provisioning vists by 37% when predator calls were played in comparison to when non-predator calls were played. They also fed offspring fewer times per visit, which resulted in 44% fewer meals in association with predator calls.


Mean number of parental provisioning visits (in one hour) in relation to whether predator (red) or non-predator (blue) calls were played. Error bars are 1 SE.

Hearing predator calls had no effect on offspring behavior – they continued to beg for food at a high rate, and did not attempt to hide.

Some parents were much more scared than others – in fact, some parents were not scared at all. The researchers measured parental fearfulness by subtracting the number of provisioning visits by parents during predator calls from the number of visits during non-predator calls. A more positive number indicated a more fearful parent (a negative number represents a parent who fed more in the presence of predator calls). The researchers discovered that more fearful parents tended to have offspring that were in poorer condition at day 6 and at fledging.


Offspring weight on day 6 (open circles) and at fledging (solid circles) in relation to parental fearfulness.  Higher positive numbers on x-axis indicate increasingly fearful parents.

Importantly, more fearful parents tended to have offspring that died at an earlier age. Based on this finding, the researchers created a statistical model that compared survival of offspring that heard predator playbacks throughout late-development with survival of offspring that heard non-predator playbacks during the same time period. They estimated a 24% reduction in survival. Combined with their previous study on playbacks during early development, the researchers estimate that hearing predator playbacks throughout early and late development would reduce offspring survival by an amazing 53%.

This “fear itself” phenomenon can extend to other trophic levels in a food web. For example recent research by Zanette and a different group of researchers showed that playbacks of large carnivore vocalizations dramatically reduced foraging by raccoons on their major prey, red rock crabs. When these carnivore playbacks were continued for a month, red rock crab populations increased sharply. This increase in crab population size was followed by a decline of the crab’s major competitor – the staghorn sculpin, and the crab’s favorite food, a Littorina periwinkle. Thus “fear itself” can cascade through the food web, affecting multiple trophic levels in important ways that ecologists are now beginning to understand.

note: the paper that describes this research is from the journal Ecology. The reference is Dudeck, B. P., Clinchy, M., Allen, M. C. and Zanette, L. Y. (2018), Fear affects parental care, which predicts juvenile survival and exacerbates the total cost of fear on demography. Ecology, 99: 127–135. Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2018 by the Ecological Society of America. All rights reserved.

Crawling with caterpillars courtesy of climate change – and ants

In the book of Exodus, Yahweh inflicts upon the Egyptians ten plagues, several of which have biological bases. Plagues two three, four and five are frogs, lice, wild animals and diseased livestock. But it is the eighth plague that is relevant to today’s tale – the locust explosion. As it turns out, insect populations have periodically exploded throughout recorded history (and no doubt before), and for many years ecologists have been trying to understand why insect populations are so variable. Rick Karban has taught a field course at Bodega Marine Reserve, California since 1985, and, as he describes “In some years, the bushes are dripping with caterpillars and in others they are very difficult to find.  The wooly bears (Platyprepia virginalis) are so conspicuous and charismatic that I couldn’t help wondering what was responsible for their large swings in abundance (they are more than 1,000 times as abundant in big years than in lean ones).”


Wooly bear caterpillar density during annual surveys conducted in march of each year.

The early stage caterpillars are most common in wet marshy habitats, but as they develop, they move to dryer upland habitats where they pupate, metamorphose into moths and mate. Young caterpillars live in leaf litter, eating vegetation and decaying organic matter.


Late instar (close to pupation) wooly bear caterpillar feeding on bush lupine. Credit Rick Karban.

Karban and his colleagues recognized that insect populations are sensitive to climate, and wondered whether climate change may be playing a role in Platyprepia population explosions. But there’s much more to climate change than global warming; for example, many areas of the world expect much more variable precipitation patterns, with more big storms and more droughts. Karban and his colleagues wanted to know whether variable precipitation might affect wooly bear populations. So they examined rainfall records between 1983 and 2016, and found that numerous heavy rainfall events (over 5 cm) in the previous year were correlated with increases in caterpillar abundance.


Change in caterpillar abundance in relation to number of heavy rainfall events (over 5 cm) during the previous year. Note the y-axis is the natural logarithm of the change in abundance.

Karban and his students explored three hypotheses for why caterpillars increased following a year with numerous heavy rainfall events. First, perhaps more rain causes more plant growth and deeper litter, providing extra food for caterpillars. Second, heavy rains may reduce the number of predacious ground-nesting ants. Lastly, heavy rains may produce deeper denser litter providing refuge from predacious ants.

The researchers tested litter as food hypothesis by comparing caterpillar growth rates during the summer, which usually has very little rainfall. They weighed individual caterpillars, placed them into cages and supplied them with litter from either wet or dry sites. After 30 days, they reweighed the caterpillars and found that all of them had lost weight, and that there were no significant differences in weight change between wet and dry sites. Thus, at least during the summer, there was no evidence that wet sites had better food for caterpillars.

Karban and his colleagues turned their attention to ants.


If ants stayed away from wet sites, that would suggest that rainy years may benefit caterpillars by reducing the number of ants in their habitat. To measure ant abundance, the researchers set out bait stations supplied with a sugar-laced cotton ball and 1 cm3 of hot dog. They discovered many more ants, and in particular, many more Formica lasioides (a fearsome caterpillar-killer) ants were recruited to dry sites than wet sites. This suggested that years with numerous rainfall events might reduce ant abundance, at least in the wet areas preferred by young caterpillars.

The researchers tested the ant predation hypotheses by caging caterpillars in plastic deli containers that had either window screen bottoms that allowed ants to enter but prevented the caterpillars from leaving, or had spun polyester bottoms that prevented ant access. At each of 12 field sites, the caterpillars were caged with litter that matched the depth and wetness of litter found at that site. All caterpillars protected from ants survived, while 40% of the unprotected caterpillars from dry sites and 23% of the unprotected caterpillars from wet sites were killed by predators. So ants are clearly fearsome predators, but more so under dry conditions.


A Formica lasioides ant subdues an early install wooly bear caterpillar within the confines of a deli box. Credit Rick Karban.

But does litter wetness help protect against predacious ants? To investigate this question, the researchers placed caterpillars in deli containers that permitted ant access. At each site, two containers were placed side-by-side; one contained a caterpillar + litter from a wet site, while the other contained a caterpillar + litter from a dry site. Both containers were completely filled with litter and left in the field for 48 hours. The researchers discovered that caterpillars were 26% more likely to avoid predation if they were in a container stuffed with litter from a wet site. This suggests that litter from wet sites acts as a refuge for caterpillars against predators.


Caterpillar survival rate in relation to litter wetness.

Unfortunately, no long-term data on ant abundance are available, so we don’t know the relationship between ant and caterpillar abundance over time. But when ants were excluded, caterpillars survived well, and when ants were present, caterpillars survived best in wet sites with deep litter. It is not clear why caterpillars survive ant predation better in wet litter. One possibility is that caterpillars are more active than ants at cooler temperatures, and may be more likely to avoid them in wet and cool conditions. A second possibility is that dry litter is structurally less complex than wet litter, and ants may be more likely to move efficiently to capture caterpillars in dry terrain.

Given the predictions for more rainfall variability in coming years, Karban and his colleagues expect caterpillar abundance to fluctuate even more dramatically from year to year. In this system, and presumably other insect populations as well, multiple factors interact to determine whether there will be a population outbreak reminiscent of Pharaoh’s experience early in recorded history.

note: the paper that describes this research is from the journal Ecology. The reference is Karban, Richard, Grof-Tisza, Patrick, and Holyoak, Marcel (2017), Wet years have more caterpillars: interacting roles of plant litter and predation by ants. Ecology, 98: 2370–2378. doi:10.1002/ecy.1917. Thanks to the Ecological Society of America for allowing me to use figures from the paper. Copyright © 2017 by the Ecological Society of America. All rights reserved.