Sharon P. Lawler

Food web architecture and population dynamics in laboratory microcosms of protists.

In theory, food chain length and omnivory are pivotal elements of food web structure that can affect the population dynamics of species within the web. Long food chains are thought to be less stable than shorter food chains, and omnivores are thought to destabilize food webs, although populations of omnivores may be more stable than populations of nonomnivores. In three of four simple food webs assembled from bacteria and protists in laboratory microcosms, the abundance of bacterivorous protists varied more over time when the species occurred in longer versus shorter food chains. The abundance of protists attacked by omnivorous top predators was either more or less temporally variable than in webs where top predators fed only at one adjacent trophic level, depending on the particular combination of interacting species. The abundance of omnivorous top predators varied less over time than the abundance of top predators restricted to feeding only at an adjacent trophic level. Observations of increased temporal variation in prey abundance in longer food chains and low temporal variation in omnivore abundance agree broadly with several predictions of food web theory. The observation that different species in similar trophic positions can exhibit very different dynamics suggests that stability may depend on complex interactions between species-specific life-history traits and general patterns of food web architecture.

The role of dispersal in predator-prey metapopulation dynamics

We report the role of dispersal in the metapopulation dynamics of a protist predator-prey pair, the predaceous ciliate Didinium nasutum Muller feeding on the bacterivorous ciliate Colpidium cf. striatum Stokes. In previous work we showed that this extinction-prone pair persisted as metapopulations in subdivided habitats. An experiment assessed the effects of habitat subdivision on persistence and dynamics. Undivided habitats were 270 or 750 mL in volume, and subdivided habitats (arrays) were sets of nine or 25 linked 30mL bottles (270 or 750mL total volume), each replicated three times. Undivided microcosms allowed maximum dispersal, whereas subdivision reduced dispersal. Within arrays, bottles with more connecting tubes allowed more dispersal. Nine and 25 bottle arrays also differed in the mean number of connections per bottle. The effects of dispersal on predator-prey dynamics were tested by comparing subdivided vs. undivided microcosms, bottles with different numbers of connecting tubes, and nine vs. 25 bottle arrays. We tested the following predictions from metapopulation theory. (i) Predator and prey persistence and predator abundance will be greatest at intermediate dispersal rates. (ii) Prey abundance, local population variability and asynchrony in population fluctuations will be greatest at low dispersal rates. (iii) Predator :prey ratios will be greatest at high dispersal rates. Predictions were confirmed, except for the following. (i) Two measures of synchrony differed in whether they showed the expected pattern. Spatial synchrony (estimated via correlation of densities among patches within sampling dates) showed high variance and did not vary with dispersal rates. However, spatial variability (CV of density across adjacent pairs of linked bottles), showed the predicted decrease with increased dispersal. (ii) Evidence that dispersal increases predator :prey ratios was inconclusive. Predator :prey ratios were lower in undivided 750mL microcosms than in 750mL arrays, possibly because predators over-exploited prey in undivided microcosms, so that both became scarce. Conversely, within arrays, predator :prey ratios were greatest in bottles that allowed the most dispersal, as predicted. This work generally confirms the predicted effects of dispersal on predator-prey metapopulation dynamics. It also demonstrates the need for models to include more realism, e.g. the possibility of over-exploitation with very high dispersal.

Competition Between Aquatic Insects and Vertebrates: Interaction Strength and Higher Order Interactions

Replicated experiments in artificial ponds demonstrated that an assemblage of aquatic insects competed with tadpoles of the frogs Hyla andersonii and Bufo woodhousei fowleri. We independently manipulated the presence or absence of aquatic insects, and the abundance of an anuran competitor (O or 150 Bufo w. fowleri per experimental pond), using a completely crossed design for two—factor variance analysis, and observed the responses of initially similar cohorts of Hyla andersonii tadpoles to neither, either, or both insect and anuran competitors. Insects and Bufo significantly depressed the mean individual mass at metamorphosis of Hyla froglets and the cumulative biomass of anurans leaving the ponds at metamorphosis. Neither insects nor Bufo affected the survival or larval period of Hyla. Insects also significantly reduced the mean mass of Bufo, showing that both anurans responded to competition from insects. The intensity of competition between natural densities of insects and Hyla tadpoles was comparable to the intensity of competition between Bufo and Hyla, as a density of 150 Bufo/1000 L.

THERMAL PHYSIOLOGY, PHENOLOGY, AND DISTRIBUTION OF TREE FROGS

Seasonal and latitudinal gradients in the thermal sensitivity of jumping performance have been examined in 10 species of hylids representing five different seasonal and distributional categories. Among temperate species, low-temperature performance is best in early-breeding northern species, intermediate in later-breeding northern species, and worst in southern species. Subtropical and tropical species have progressively higher critical thermal minima, but their thermal performance curves are not much different from those of southern temperate species in other respects. Critical upper thermal limits on jumping performance are much less variable than are lower limits, and the critical thermal minimum (CTmin) is not correlated with the critical thermal maximum (CTmax) in our sample. Effectively, CTmax is relatively inflexible, and consequently an improvement in low-temperature performance (reduction in L80, the lowest temperature at which frogs can attain 80% of maximal performance) results in a widening of the thermal breadth of performance at or above some arbitrarily selected standard (e.g., B80, 80% of maximal performance). The strong negative correlation between L80 and B80 in our sample of hylids is not found in lizards. Differences in thermal sensitivity of jumping performance are strongly correlated with naturally experienced minimum body temperatures. The thermal sensitivity of locomotion may itself be (or may be correlated with) a mechanism underlying distributional differences and temporal partitioning of habitat among frogs. We suggest that some aspects of the community organization of hylids, such as breeding phenology and the depauperate faunas of high latitudes, are constrained at least in part by thermal physiology. Ecological consequences of this constraint are currently being investigated.

Ecology and Breeding Phenology of Larval Hyla Andersonii: The Disadvantages of Breeding Late

We exploited natural variation in the breeding phenology of Hyla andersonii, the Pine Barrens Treefrog, to test whether offspring hatching at different times during the breeding season differed in growth and survival. We studied the growth and survival of two successive cohorts of Hyla tadpoles in 12 artificial ponds, where we also manipulated abundances of two kinds of organisms that might interact seasonally with Hyla: aquatic insects and Bufo tadpoles. We also measured whether the abundance of periphyton changed seasonally in response to manipulations of aquatic insects and tadpoles, to describe temporal patterns of resource availability that might affect tadpole growth and survival. Additions of hatchlings occurred 34 d apart, and the first cohort completed development before the addition of the second cohort. Additions of aquatic insects and Bufo woodhousii tadpoles, separately or together, did not significantly affect survival in either cohort. Although the initial larval density of the first cohort was 1.5 times that of the second, the first cohort survived better, grew and developed more rapidly, and metamorphosed at larger size than the second cohort. Com- petition from Bufo tadpoles and insects, as inferred from reduced growth rates relative to controls without these interspecific competitors, was more pronounced in the first cohort than in the second, and coincided with reduced standing crops of periphyton, an important food for tadpoles. Although breeding late in the season reduced the growth and survival of offspring, potential reductions in the fitness of late-breeding adults are problematic because of un- certainties about the number of breeding efforts made by each female each year. Breeding late would be disadvantageous if different subpopulations breed only early or only late in the season, and if the patterns seen in our study hold for natural ponds. However, if late breeding efforts represent additional clutches deposited by adults that have already repro- duced earlier that year, production of additional offspring later in the season could increase fitness.

Effects of Basal Resources, Predation, and Alternative Prey in Microcosm Food Chains

Ecological theorists propose that the species composition of trophic levels can influence the relative strengths of top-down (predation) or bottom-up (nutrient) effects in food chains. We tested this by constructing two- and three-level food chains of bacteria and protists. Bacteria made up the first trophic level. The second level contained Chilomonas paramecium alone, Colpidium cf. striatum alone, or both together. Predatory Euplotes patella occupied the third trophic level. These assemblages were cultured in microcosms containing either low- or high-nutrient medium. Manipulating nutrients and predation produced comparable changes in the abundance of bacteria. In the second trophic level, Chilomonas was usually driven extinct by predation, but Colpidium was affected more by nutrients than predation because it had a partial size refuge from predation. Both species survived more poorly with predatory Euplotes if the other was present, because the predator was more abundant and persistent when both prey were available. The predator was less persistent in high-nutrient microcosms, because additional nutrients boosted the proportion of Colpidium populations within the size refuge. This type of mechanism could limit trophic cascades in food chains where resources affect prey vulnerability.

Species richness, species composition and population dynamics of protists in experimental microcosms

Laboratory experiments using communities assembled from bacteria plus one, two or four protist predator-prey pairs tested whether species richness affects the persistence, abundance and temporal variability of the protist species. Comparisons among six different food webs of four protist species tested whether the effects of species composition rival the effects of species richness. Most populations in species-rich food webs were less abundant than populations of the same species in food webs with fewer species. Increased species richness increased the variability of one species. Combining predator- prey pairs increased the number of extinctions. The species composition of food webs also affected the persistence and mean abundances of many species, and changed the variability of two species

Transmission of Ehrlichia risticii, the agent of Potomac horse fever, using naturally infected aquatic insects and helminth vectors: preliminary report.

Ehrlichia risticii, the agent of Potomac horse fever (PHF), has been recently detected in trematode stages found in snail secretions and in aquatic insects. Based on these findings, horses could conceivably be exposed to E. risticii by skin penetration with infected cercariae, by ingestion of infected cercariae in water or via metacercariae in a second intermediate host, such as an aquatic insect. In order to test this hypothesis, horses were challenged with infectious snail secretions and aquatic insects collected from a PHF endemic region in northern California. Two horses stood with their front feet in water harbouring E. risticii-infected cercariae, 2 horses drank water harbouring E. risticii-infected cercariae, and 6 horses were fed pools of different aquatic insects harbouring E. risticii-infected metacercariae. In this preliminary study, only the one horse infected orally with mature caddisflies (Dicosmoecus gilvipes) developed the clinical and haematological disease syndrome of PHF. The agent was isolated from the blood of the infected horse in a continuous cell line and identified as E. risticii by characterisation of the 16S rRNA gene. Therefore, E. risticii is maintained in nature in a complex aquatic ecosystem and transmission to horses can occur through accidental ingestion of insects such as caddisflies containing infected metacercariae. At present, the small number of horses used in this study does not exclude other insects and free trematode stages as potential sources of infection.

Effects of Food Chain Length and Omnivory on Population Dynamics in Experimental Food Webs

Food web theory predicts that different patterns of trophic connections will have different consequences for the population dynamics of individual species (MacArthur, 1955), as well as the system dynamics of larger multispecies communities (May, 1973; Pimm and Lawton, 1977). Despite the central role that population-dynamic models have played in the development of food web theory, we still know astonishingly little about whether the long-term population dynamics of real species vary among food webs of different structure in the ways that theory predicts (Lawton, 1989). One way to address this critical dearth of information is to build simple food webs of known structure, and then observe the dynamics of the species that they contain (Lawler and Morin, 1993). This experimental approach is most tractable in laboratory microcosms where the long-term dynamics of species with short generation times can be observed readily over short time periods. Here we extend our previous studies of protist population dynamics in simple food webs of known structure to explore some predicted relations between food web structure and population dynamics. Such studies are absolutely essential to assess whether food web theory makes reasonable predictions about population dynamics.

Infection of aquatic insects with trematode metacercariae carrying Ehrlichia risticii, the cause of Potomac horse fever.

Abstract We provide evidence ofEhrlichia risticiiHolland, the agent of Potomac horse fever, in trematode stages found in aquatic insects collected from a pasture stream in northern California, using nested polymerase chain reaction (PCR) amplification and sequence analyses of the 16S rRNA, 51 kDa major antigen andgroELheat shock protein genes.E. risticiiwas detected in metacercariae found in the immatures and adults of the following insects: caddisflies (Trichoptera), mayflies (Ephemeroptera), damselflies (Odonata, Zygoptera), dragonflies (Odonata, Anisoptera), and stoneflies (Plecoptera). The prevalence ofE. risticiiwas 31.9% (n= 454 individuals) in aquatic insects (13 of 17 species were positive). Prevalence within orders was as follows: 43.5% (n= 207) in caddisflies, 15.2% (n= 92) in mayflies, 13.9% (n= 115) in damselflies, 10.0% (n= 10) in dragonflies, and 80.0% (n= 30) in stoneflies. This study demonstrates a broad intermediate host range for trematodes that act as vector forE. risticii.Insects are likely to play an important role in the epidemiology of this disease.