Locally abundant, endangered Mariana swiftlets impact the abundance, behavior, and body condition of an invasive predator.

Invasive predators are known to have negative consumptive and non-consumptive effects on native species, but few examples show how the abundance of native prey may influence an established invasive predator. We compared invasive brown treesnakes (Boiga irregularis; BTS) found in caves occupied by endangered Mariana swiftlets (Aerodramus bartschi) to snakes found in nearby forests and caves without birds to quantify how the abundance of native avian prey impacts BTS abundance and behavior on Guam. From 2011 to 2017 we removed 151 BTS in caves occupied by swiftlets and never observed BTS in caves without birds. Notable locations included snakes foraging near swiftlets and in holes that allowed cave access and escape from capture. Of 43 BTS with gut contents, 27 (63%) contained swiftlets. BTS in swiftlet-occupied caves had greater fat mass compared to forests, indicating access to swiftlets may increase body condition and promote reproduction. Number of ovarian follicles was significantly greater in female snakes from swiftlet-occupied caves compared to those from ravine, but not limestone forests; evidence of male BTS being more capable of reproduction was limited (i.e., fewer non-discernible but not significantly larger testes in snakes from caves). Assuming other limiting factors are considered, altering the functional response of predators through the modification of caves or interdiction lures to exclude or hinder the largest BTS could bolster swiftlet populations by increasing nesting refugia in currently-occupied caves and facilitate recolonization of historical caves.

In vitro comparison of treatments and commercially available solutions on mortality of Angiostrongylus cantonensis third-stage larvae.

On Hawai'i Island, an increase in human neuroangiostrongyliasis cases has been primarily associated with the accidental ingestion of Angiostrongylus cantonensis L3 in snails or slugs, or potentially, from larvae left behind in the slug's slime or feces. We evaluated more than 40 different treatments in vitro for their ability to kill A. cantonensis larvae with the goal of identifying a safe and effective fruit and vegetable wash in order to reduce the risk of exposure. Our evaluation of treatment lethality was carried out in two phases; initially using motility as an indicator of larval survival after treatment, followed by the development and application of a propidium iodide staining assay to document larval mortality. Treatments tested included common household products, consumer vegetable washes and agricultural crop washes. We found minimal larvicidal efficacy among consumer-grade fruit and vegetable washes, nor among botanical extracts such as those from ginger or garlic, nor acid solutions such as vinegar. Alkaline solutions, on the other hand, as well as oxidizers such as bleach and chlorine dioxide, did show larvicidal potential. Surfactants, a frequent ingredient in detergents that lowers surface tension, had variable results, but dodecylbenzene sulfonic acid as a 70% w/w solution in 2-propanol was very effective, both in terms of the speed and the thoroughness with which it killed A. cantonensis L3 nematodes. Thus, our results suggest promising directions for future investigation.

Vertical foraging shifts in Hawaiian forest birds in response to invasive rat removal.

Worldwide, native species increasingly contend with the interacting stressors of habitat fragmentation and invasive species, yet their combined effects have rarely been examined. Direct negative effects of invasive omnivores are well documented, but the indirect effects of resource competition or those caused by predator avoidance are unknown. Here we isolated and examined the independent and interactive effects of invasive omnivorous Black rats (Rattus rattus) and forest fragment size on the interactions between avian predators and their arthropod prey. Our study examines whether invasive omnivores and ecosystem fragment size impact: 1) the vertical distribution of arthropod species composition and abundance, and 2) the vertical profile of foraging behaviors of five native and two non-native bird species found in our study system. We predicted that the reduced edge effects and greater structural complexity and canopy height of larger fragments would limit the total and proportional habitat space frequented by rats and thus limit their impact on both arthropod biomass and birds' foraging behavior. We experimentally removed invasive omnivorous Black rats across a 100-fold (0.1 to 12 ha) size gradient of forest fragments on Hawai'i Island, and paired foraging observations of forest passerines with arthropod sampling in the 16 rat-removed and 18 control fragments. Rat removal was associated with shifts in the vertical distribution of arthropod biomass, irrespective of fragment size. Bird foraging behavior mirrored this shift, and the impact of rat removal was greater for birds that primarily eat fruit and insects compared with those that consume nectar. Evidence from this model study system indicates that invasive rats indirectly alter the feeding behavior of native birds, and consequently impact multiple trophic levels. This study suggests that native species can modify their foraging behavior in response to invasive species removal and presumably arrival through behavioral plasticity.

Brown tree snake (Boiga irregularis) population density and carcass locations following exposure to acetaminophen.

Mass aerial delivery of dead mouse baits treated with acetaminophen has been evaluated as a means to reduce brown tree snake (Boiga irregularis) populations over large areas, increasing the likelihood of wide-scale eradication on Guam. Given the high density of snakes in some areas of their invasive range, eradication efforts could result in a resource pulse that may influence food web dynamics and the indirect transport of acetaminophen among trophic levels. We evaluated abundance, habitat type, and snake size (i.e., age) within two study sites on Guam, a secondary limestone forest (upland) and an abandoned coconut plantation (coastal), to determine how experimentally dosing snakes with acetaminophen is likely to influence carrion availability. We found snakes trapped in 3.24 ha plots occurred in greater abundance (population size = 72.5 snakes; SE = 8.8) and were significantly larger (978.6 mm, SE = 14.9) in the coastal than in the upland site (population size = 26.9, SE = 21.5; length = 903.0 mm, SE = 15.9). Despite these differences, carcasses of snakes that died after consuming acetaminophen-laced mice (80 mg) were recovered in consistent locations between sites, with 92 % located on the ground, 4 % in trees, and 4 % found in rock cavities at both sites. Given that most snakes were found on the ground rather than in the tree canopy, our results suggest that many poisoned snake carcasses will be accessible to a wide range of potential scavengers, possibly influencing food web dynamics and potentially contributing to indirect toxicant transfer within affected ecosystems.

Detection of Angiostrongylus cantonensis in the Blood and Peripheral Tissues of Wild Hawaiian Rats (Rattus rattus) by a Quantitative PCR (qPCR) Assay.

The nematode Angiostrongylus cantonensis is a rat lungworm, a zoonotic pathogen that causes human eosinophilic meningitis and ocular angiostrongyliasis characteristic of rat lungworm (RLW) disease. Definitive diagnosis is made by finding and identifying A. cantonensis larvae in the cerebral spinal fluid or by using a custom immunological or molecular test. This study was conducted to determine if genomic DNA from A. cantonensis is detectable by qPCR in the blood or tissues of experimentally infected rats. F1 offspring from wild rats were subjected to experimental infection with RLW larvae isolated from slugs, then blood or tissue samples were collected over multiple time points. Blood samples were collected from 21 rats throughout the course of two trials (15 rats in Trial I, and 6 rats in Trial II). In addition to a control group, each trial had two treatment groups: the rats in the low dose (LD) group were infected by approximately 10 larvae and the rats in the high dose (HD) group were infected with approximately 50 larvae. In Trial I, parasite DNA was detected in cardiac bleed samples from five of five LD rats and five of five HD rats at six weeks post-infection (PI), and three of five LD rats and five of five HD rats from tail tissue. In Trial II, parasite DNA was detected in peripheral blood samples from one of two HD rats at 53 minutes PI, one of two LD rats at 1.5 hours PI, one of two HD rats at 18 hours PI, one of two LD rats at five weeks PI and two of two at six weeks PI, and two of two HD rats at weeks five and six PI. These data demonstrate that parasite DNA can be detected in peripheral blood at various time points throughout RLW infection in rats.

Capillaria hepatica infection in black rats (Rattus rattus) on Diego Garcia, British Indian Ocean Territory.

Rats (Rattus spp.) are among the most damaging invasive species worldwide. The accidental introduction of rats has caused significant detriment to native flora and fauna, crops, structures, and human livelihoods. Rats are vectors of disease and carriers of various zoonotic parasites. Capillaria hepatica (syn. Callodium hepaticum) is a parasitic nematode found primarily in rodents but is known to infect over 140 mammal species, including human beings and several species of domestic animals. In this case study, the presence of C. hepatica infection in black rats on Diego Garcia, British Indian Ocean Territory, is reported. Liver samples from 20 black rats (Rattus rattus) were collected during a concurrent population density estimation study. Histology revealed 15 (75%) of the rats sampled had a current or previous infection with C. hepatica. In addition, a larval cestode compatible in size and shape with Cysticercus fasciolaris, the larval stage of Taenia taeniaeformis of cats, was found in 3 (15%) of the rats sampled. The high prevalence of C. hepatica infection in rats on Diego Garcia has implications for human health given the high population density of rats found on the island.

Potential citric acid exposure and toxicity to Hawaiian hoary bats (Lasiurus cinereus semotus) associated with Eleutherodactylus frog control.

We examined potential exposure of Hawaiian hoary bats (Lasiurus cinereus semotus) to citric acid, a minimum risk pesticide registered for control of invasive Eleutherodactylus frog populations. Hoary bats are nocturnal insectivores that roost solitarily in foliage, federally listed as endangered, and are endemic to Hawaii. Oral ingestion during grooming of contaminated fur appears to be the principal route by which these bats might be exposed to citric acid. We made assessments of oral toxicity, citric acid consumption, retention of material on fur, and grooming using big brown bats (Eptesicus fuscus) as a surrogate species. We evaluated both ground application and aerial application of 16 % solutions of citric acid during frog control operations. Absorbent bat effigies exposed to ground and aerial operational spray applications retained means of 1.54 and 0.02 g, respectively, of dry citric acid, although retention by the effigies was much higher than bat carcasses drenched in citric acid solutions. A high dose delivered orally (2,811 mg/kg) was toxic to the big brown bats and emesis occurred in 1 bat dosed as low as the 759 mg/kg level. No effect was observed with the lower doses examined (≤ 542 mg/kg). Bats sprayed with 5 ml of 16 % (w/w) citric acid solution showed no evidence of intoxication. In field situations, it is unlikely that bats would be sprayed directly or ingest much citric acid retained by fur. Based on our observations, we believe Hawaiian hoary bats to be at very low risk from harmful exposure to a toxic dose of citric acid during frog control operations.

Longevity of rodenticide bait pellets in a tropical environment following a rat eradication program.

Invasive rodents (primarily Rattus spp.) are responsible for loss of biodiversity in island ecosystems worldwide. Large-scale rodenticide applications are typically used to eradicate rats and restore ecological communities. In tropical ecosystems, environmental conditions rapidly degrade baits and competition for baits by non-target animals can result in eradication failure. Our objective was to evaluate persistence of rodenticide baits during a rat eradication program on Palmyra Atoll; a remote tropical atoll with intense competition for resources by land crabs. Following aerial application, bait condition was monitored in four terrestrial environments and in the canopy foliage of coconut palms. Ten circular PVC hoops were fixed in place in each of Palmyra's four primary terrestrial habitats and five rodenticide pellets were placed in each hoop. Five coconut palms were selected in three distinct regions of the atoll. One rodenticide pellet was placed on each of five palm fronds in each coconut palm. Fresh baits were placed in all monitoring locations after each broadcast bait application. Bait condition and survival was monitored for 7 days after the first bait application and 6 days after second application. Bait survival curves differed between applications at most monitoring sites, suggesting a decrease in overall rat activity as a result of rodenticide treatment. One terrestrial site showed near 100 % bait survival after both applications, likely due to low localized rat and crab densities. Median days to pellet disappearance were one and two days for the first and second application, respectively. Differences in survival curves were not detected in canopy sites between bait applications. Median days to pellet disappearance in canopy sites were 2 and 4 days for the first and second application, respectively. Frequent rainfall likely contributed to rapid degradation of bait pellets in coconut palm fronds.

Assessing spatial variation and overall density of aerially broadcast toxic bait during a rat eradication on Palmyra Atoll.

Baits containing brodifacoum rodenticide were aerially applied to eradicate invasive black rats from Palmyra Atoll, an important biodiversity center. Bait application must be sufficient to be effective, while minimizing environmental hazards by not exceeding designated label rates, prompting our bait density assessments for two aerial drops. With few physical or human resources on this remote, uninhabited atoll, assessments were particularly challenging, requiring observations within 30 min of aerial application to avoid bait loss to rats, crabs, or elements. We estimated bait density using quadrat sampling within 13 terrestrial sampling areas. We also sampled 10 tidal flat areas to assess inadvertent bait scatter into marine aquatic environments. Of particular value for challenging sampling circumstances, our quadrats had to be lightweight and durable, which we addressed by using widely available PVC hoops ("Hula Hoops"), the size of which was ideal for sampling purposes. At 77.5 and 78.7 kg/ha, overall bait densities were very near to the target densities of 80 and 75 kg/ha, respectively. However, considerable variability in bait densities existed among sampled areas, 8.6-178.2 and 31.4-129.5 kg/ha for the respective drops, respectively. Environmental, human, and equipment factors likely accounted for this variability. Tidal flat sampling revealed variable bait scatter into aquatic environments, from 0-46.3 kg/ha across the two drops. No differences were found in average bait densities among 1-, 4-, and 7-m distances from high tide lines. Our methods might broadly assist bait density (and other) surveys under challenging circumstances.

The effect of cooking on diphacinone residues related to human consumption of feral pig tissues.

We examined feral pig tissues to determine whether the potential hazard of consuming meat from pigs previously exposed to diphacinone rodenticide baits was reduced by cooking. Residue levels were measured in cooked and uncooked tissues of feral pigs exposed to sub-lethal quantities of the anticoagulant rodenticide. Pigs were provided large amounts of baits or untreated food to consume, then euthanized prior to the onset of symptoms indicative of rodenticide poisoning or sickness. For analysis, we grouped pigs into three levels of mean diphacinone consumption: 0, 3.5, and 7.4 mg/kg. None of the pigs displayed obvious signs of toxicity during the study period. The highest concentrations of diphacinone were found in liver tissue. Cooking had little effect on residual diphacinone concentrations. The hazards to humans and pets from meat from feral pigs that consumed the rodenticide diphacinone are not reduced by cooking; consumption of pig meat obtained from areas with active rodent control programs should be avoided.

Efficacy of rodenticide baits for the control of three invasive rodent species in Hawaii.

We tested the efficacy and palatability of nine commercial rodenticide bait formulations on Polynesian rats (Rattus exulans), roof rats (R. rattus), and house mice (Mus musculus). Efficacy varied by rodenticide tested and rodent species. Generally, rodenticides were more effective against mice than for either of the rat species, and mice tended to consume more rodenticide bait than the laboratory chow alternative food. Efficacy was generally highest for the second-generation anticoagulants tested; however, this varied across products and one-first-generation rodenticide had similar effectiveness. Bait acceptance (palatability) also varied both by rodenticide and by rodent species. Acceptance was the lowest for the acute rodenticides. Bait acceptance appeared to substantially affect the efficacy of rodenticides; materials that were not well accepted produced lower mortality rates. Rodenticide products currently registered for use in Hawaii performed less effectively in this study than other available products not yet registered. Although markets for rodent control products for use on islands are limited, there are advantages to having additional products registered for island use in agriculture, conservation, and public health.

Isolation of microsatellite loci from the coqui frog, Eleutherodactylus coqui.

Thirteen microsatellite loci were isolated from the coqui frog (Eleutherodactylus coqui) and optimized for future research. The loci were screened across 37 individuals from two Puerto Rican populations. Loci were variable with the number of alleles per locus ranging from three to 38. Polymorphic information content ranged from 0.453 to 0.963 and observed heterozygosity for each population ranged from 0.320 to 0.920.

Determination of diphacinone residues in Hawaiian invertebrates.

A reversed-phase ion-pair liquid chromatographic analysis combined with a solid-phase extraction clean-up method is used to assess the quantity of diphacinone residue found in invertebrates. Three invertebrate species are exposed to commercially available diphacinone-fortified bait used for rat control. The invertebrate samples are collected, frozen, and shipped to the laboratory. The samples are homogenized after cryogenic freezing. A portion of the homogenized samples are extracted with acidified chloroform-acetone, followed by cleanup with a silica solid-phase extraction column. Diphacinone is detected by UV absorption at 325 nm after separation by the chromatographic system. The method limit of detection (MLOD) for snail and slug samples averaged 0.055 and 0.066 mg/kg, respectively. Diphacinone residues in snail tissue ranges from 0.83 to 2.5 mg/kg for Oxychilus spp. The mean recoveries from snails at 0.20 and 2.0 are 97 +/- 21% and 84 +/- 6%. Diphacinone residues in slug tissue ranges from 1.3 to 4.0 mg/kg for Deroceras laeve and < MLOD to 1.8 mg/kg for Limax maximus, respectively. The mean recoveries from slugs at 0.20 and 2.0 mg/kg are 91% +/- 15% and 86% +/- 5%.

Probabilistic risk assessment for snails, slugs, and endangered honeycreepers in diphacinone rodenticide baited areas on Hawaii, USA.

Three probabilistic models were developed for characterizing the risk of mortality and subacute coagulopathy to Poouli, an endangered nontarget avian species, in broadcast diphacinone-baited areas on Hawaii, USA. For single-day exposure, the risk of Poouli mortality approaches 0. For 5-d exposure, the mean probability of mortality increased to 3% for adult and 8% for juvenile Poouli populations. For Poouli that consume snails containing diphacinone residues for 14 d, the model predicted increased levels of coagulopathy for 0.42 and 11% of adult and juvenile Poouli populations, respectively. Worst-case deterministic risk characterizations predicted acceptable levels of risk for nonthreatened or endangered species such as northern bobwhite quail and mallards. Also, no acute toxicity was noted for snails and slugs that feed on diphacinone baits.

Coyotes (Canis latrans) are definitive hosts of Neospora caninum.

Four captive-raised coyote pups consumed tissues from Neospora caninum-infected calves. Faeces were examined from 4 days before to 28 days after infection. One pup shed N. caninum-like oocysts, which tested positive for N. caninum and negative for Hammondia heydorni using PCR tests. Coyotes are the second discovered definitive host of N. caninum, after dogs. In North America, the expanding coyote ranges and population increase the probability of contact with domestic livestock. To reduce the risk of transmission of N. caninum to intensively farmed cattle, we recommend protection of feedstuffs using canid-proof fences, and careful disposal of dead stock.