Wireworm populations in relation to crop production: I. A large-scale flotation method for extracting wireworms from soil samples and results from a survey of 600 fields

The wireworm survey in the Eastern Counties revealed many cases where the observed wireworm damage failed to correspond with the estimated field population. A possible explanation for this was the inaccuracy of counts made by picking wireworms out of the soil samples by hand. Tests showed that such methods recovered an extremely variable proportion of the wireworms in the- samples and, on the average, only two-fifths of the larvae were obtained. A modified form of the washing and flotation technique used by Salt & Hollick (1944) was introduced for large-scale work and is described. By this method, ten samples of soil (4 in. diam. and 6 in. deep) bulked together are examined at a time and can be dealt with at the rate of 13 samples per man per hour with an efficiency of 95-100% in the extraction of wireworms. The populations estimated on 600 fields sampled between December 1942 and May 1943 have thrown more light on the size and composition of the wireworm population in grass and arable fields. Inspection of the crop results on fields tested by the washing process showed a much closer relationship between wireworm population and wireworm damage than had been obtained by the hand-sorting method in the previous year.     

Studies of wireworm populations

A method has been developed, and is here briefly described, by means of which all the wireworm larvae of all stages can be collected from soil samples. By use of that method, complete wireworm populations have been obtained from soil samples of three types, two of which provide large homogeneous populations for detailed study while the third gives information about the seasonal and spatial infestation of fields. These collections show that the wireworm population of English pasture land is on the average about three times as large as has been commonly supposed. Populations have been found ranging up to ten millions per acre in the top 12 in. of soil. In two pastures studied intensively throughout the year, the wireworm population has been found to consist of large numbers of small larvae, decreasing numbers of larger larvae, and comparatively few of the very large larvae that have usually been allowed to represent the population. Such a composition is shown to be characteristic of wireworm populations under old grass at all seasons of the year and in several fields in different parts of the country. This result suggests that in research on the wireworm problem it is desirable to take into consideration the whole wireworm population throughout the year, not merely the large larvae during the cropping season.     

Food habits of pasture wireworm,Conoderus exsul (Coleoptera: Elateridae)

Abstract

The food preferences and rates of growth of the omnivorous pasture wireworm, Conoderus exsul (Sharp) were studied. Larvae of soil-living insects were preferred to germinating maize seed, and wireworm growth rate was greater on insect compared with plant food. In glasshouse pot trials, wireworms increased their attack rate as the density of grass grub (Costelytra zealandica (White)) or white-fringed weevil (Graphthognathus leucoloma Boheman) larvae was increased. In field experiments that compared grass grub numbers in cages accessible to wireworms or protected from predators, predation varied with season from a low of 15% in late summer up to 82% in late autumn. White-fringed weevil larval mortality was increased by addition of wireworms in two field experiments, but not affected in a third trial where initial white-fringed weevil density was low. Grass grub mortality was increased by the presence of wireworms in field plots. Adult C. exsul were observed feeding on sweetcorn pollen.     

WIREWORM POPULATIONS IN RELATION TO CROP PRODUCTION: III. POPULATION CHANGES AFTER SUMMER PLOUGHING

Six permanent pastures broken up in the summer of 1943 were sampled once a fortnight after ploughing in order to study changes in the wireworm populations.
The graphs of the running means of four consecutive counts showed the chief trends in population. On four of the fields, a slight increase in population was apparent up to December. Only in February and March was a definite decline apparent. On the other two fields, the counts declined by about 30 % during the autumn and winter.
A marked fall in population (from 30 to 70%) was evident on all the fields only after the first crop had been harvested. The new population at this time was characterized by an absence of larvae less than 4 mm. in length.
The results considered alone indicate that after summer ploughing the wireworm population remains unaffected until the following spring and summer during the cropping period. When the results are considered in relation to the grass 'controls', it appears that a net decline in population occurs during the autumn and winter and that the decline in spring and summer is not more marked than that which normally occurs at this time in grass fields. Unlike the grass fields, however, the reduction is permanent.     

WIREWORM POPULATIONS IN RELATION TO CROP PRODUCTION: II. POPULATION CHANGES IN GRASSLAND

Six grass fields were sampled once a fortnight from July 1943 to September 1944, and the wireworms in the samples counted by the washing and flotation method.
The counts from twenty standard samples showed huge fluctuations which rendered them practically valueless as estimates of wireworm populations when treated singly. When treated as running means of four consecutive samplings, the counts showed certain seasonal trends, with minimum populations in July and August, rising populations throughout the autumn, maximum populations in the months from January to April and a sharp decline to the minimum populations again from April to July. On the average, the counts taken in winter were twice as high as those taken in the summer months.
The fluctuation was evident in all size groups based on larval length but was most marked in the wireworms 'under 4 mm.', '4 mm.' and '5 mm.'
Samples taken to depths of 12 and 24 in. showed that on the average about 75 % of the wireworms were found in the 0–6 in. layer and about 90% in the 0–12 in. layer. These proportions varied considerably with the seasons.
The seasonal changes in wireworm populations observed do not correspond with those expected from the life history of Agriotes spp. Possible explanations for this anomaly are discussed, but it cannot be attributed to any known technical or biological factor. Several possibilities remain to be explored.     

An investigation inot the use of multiple site characteristics to predict the presence and infestation level of wireworms (Agriotes sup., Coleoptera: Elateridae) in individual grass fields

Sixty-two grass fields were sampled in England and Wales over a three year period to assess the level of wireworm population present. Data on site-specific variables including soil physical characteristics, grass duration, grass genera diversity and other abiotic factors such as field aspect, altitude, and meteorological parameters were abiotic factors such as field aspect, altitude, and meteorological parameters were collected at each site. Only grass duration and soil bulk density showed any association with wireworm infestation stauts when considered as single variates. Data from a sub-set of 41 fields were used to develop a series of multi-variate discriminant rules to predit wireworm presence/absece in individual fields. These were vaildated using data from the remaining 21 fields. In general, the rules tended to over-estimate the number of wireworm-infested fields by misclassifying uninfested fields as infestes. Multivariate models to predict wireworm populaiton levels in infested fields were also developed using multiple and generalised linear regression. The predictive accuracy of these was poor. Neither the population prediciton models nor the presence/absence rulses accounted fully for the large inte-field variation in wireworm infestation status.     

Food Preference of Wireworms Analyzed with Multinomial Logit Models

Many species of wireworms (larvae of click beetles, Elateridae) are poly- phagous root herbivores. In grasslands under restoration succession with various grass species, we aim to determine the role of wireworms in aboveground vegetation succession. Therefore, it is crucial to know whether wireworms prefer some food plants to others. We have investigated the root preference to different grass species for Agriotes obscurus and Athous haemorrhoidalis and whether these preferences can be explained by covariates. In Experiment 1, individual wireworms could choose between four different plants, one of each species (Anthoxanthum odoratum, Festuca rubra, Holcus lanatus, and Lolium perenne). In Experiment 2, groups of wireworms were released into the soil in the center of 16 plants (4 from each species). We used multinomial logit models (MLMs) to analze the data. In the appendix the use of multinomial response models is clarified with a fictitious example, using the SAS statistic software package. No preference was found in Experiment 1. In Experiment 2 we found differences in attractiveness of plant species depending on wireworm species: A. obscurus preferred grass species from nutrient-rich grasslands (L. perenne and H. lanatus). Both wireworm species disliked F. rubra. The distance from the release point influenced the probability of being found at a certain place at the end of the experiment: wireworms tended to stay in the proximity of the release point. A. haemorrhoidalis was more often found farther from the point of release than A. obscurus. Dispersal was farther from the release point in experiments with young plants (6 weeks) compared to older ones (9 weeks). Results are discussed in a broad ecological context.     

Aggregation and mortality of Agriotes obscurus (Coleoptera: Elateridae) at insecticide-treated trap crops of wheat

Agriotes obscurus (L.) wireworms assembled in increasing numbers at rows of treated (Agrox DL Plus seed treatment) and untreated wheat, Triticum aestivum L., planted at increasing densities (0, 0.15, 0.30, and 0.60 seeds/cm). In treated wheat plots at all planting densities, no wireworm damage to seedlings was apparent, and total wireworms taken in core samples in wheat rows increased according to the asymptotic equation y = B0(1 -e(-Blx)), where B0 is the asymptote, B1 is the slope of the initial rise, and x is the seeding density. The number of dead wireworms in treated plots increased linearly and intercepted the asymptotic models (theoretical point at which 100% mortality of assembled population occurs) at 0.95 seeds/cm on 11 June and 1.14 seeds/cm on 18 June 1996. Untreated wheat at all densities planted had severe wireworm damage and significantly reduced stand. Populations that had assembled at the surviving untreated wheat were fewer than in the treated wheat plots, and although increasing with seeding density, did not follow the asymptotic model. The data suggest that A. obscurus populations can be assembled and killed in fallowed fields in large numbers at treated trap crops of wheat over a 19-d period when planted in rows spaced 1 m apart at a linear seeding density of 1.5 seeds/cm.     

Influence of preceding crop on wireworm (Coleoptera: Elateridae) abundance in the coastal plain of North Carolina

Three studies were conducted to determine the effect of preceding crop on wireworm (Coleoptera: Elateridae) abundance in the coastal plain of North Carolina. In all three studies, samples of wireworm populations were taken from the soil by using oat, Avena sativa L., baits. Treatments were defined by the previous year's crop and were chosen to reflect common crop rotations in the region. Across all three studies, eight wireworm species were recovered from the baits: Conoderus amplicollis (Gyllenhal), Conoderus bellus (Say), Conoderus falli (Lane), Conoderus lividus (Degeer), Conoderus scissus (Schaeffer), Conoderus vespertinus (F.), Glyphonyx bimarginatus (Schaeffer), and Melanotus communis (Gyllenhal). The effect of corn, Zea mays L.; cotton, Gossypium hirsutum L.; fallow; soybean, Clycine max (L.) Merr.; sweet potato, Ipomoea batatas (L.) Lam.; and tobacco (Nicotiana spp.) was evaluated in a small-plot replicated study. M. communis was the most frequently collected species in the small-plot study and was found in significantly higher numbers following soybean and corn. The mean total number of wireworms per bait (all species) was highest following soybean. A second study conducted in late fall and early spring assessed the abundance of overwintering wireworm populations in commercial fields planted to corn, cotton, peanut (Arachis hypogaea L.), soybean, sweet potato, and tobacco in the most recent previous growing season. C. lividus was the most abundant species, and the mean total number of wireworms was highest following corn and soybean. A survey was conducted in commercial sweet potato in late spring and early summer in fields that had been planted to corn, cotton, cucurbit (Cucurbita pepo L.), peanut, soybean, sweet potato, or tobacco in the most recent previous growing season. C. vespertinus was the most abundant species, and the mean total number of wireworms per bait was highest following corn.     

Evaluation of trap crops for the management of wireworms in spring wheat in Montana

The polyphagous larvae of click beetles (Coleoptera: Elateridae) are major pests of spring wheat in Montana, USA. Presently available insecticides are unable to provide control over wireworm populations, and the use of natural enemies has not been successful under field conditions. In this study, we examined the effect of seven trap crops: pea, lentil, canola, corn, durum, barley, and wheat, for their attractiveness to wireworms compared to spring wheat. Experimental plots were located in two commercial grain fields in Valier and Ledger, Montana, USA and the trials took place from May to August in 2015 and 2016. Wheat plants damaged by wireworms were recorded and their relative locations in wheat rows and adjacent trap crop rows within a plot were determined using destructive soil samples. In 2016, variable row spacing (0.25, 0.5, 0.75, and 1 m) between the trap crops (pea and lentil) and wheat was assessed. Shade house bioassays were conducted using potted pea, lentil, and wheat plants to support field trial results. Limonius californicus larvae, released at the center of each pot were sampled 4 and 10 days after sowing. Wheat intercropped with pea and lentil had significantly fewer damaged wheat plants. Wireworm numbers were lower in wheat intercropped with pea compared to the control for both locations and years. Shade house results corresponded with field results, with more wireworms collected from pea and lentil than wheat. In the spacing trials, wheat plant counts were also significantly higher when paired with pea and lentil, particularly at 0.5 m spacing. Regardless of inter-row spacing, significantly fewer wireworms were associated with wheat when intercropped with pea and lentil trap crops.     

Varietal susceptibility of potatoes to wireworm herbivory

1 Wireworms, the soil dwelling larvae of click beetles, Agriotes spp., have recently become a more prevalent pest of potatoes. The present study investigated whether potato varieties showed variable susceptibility to wireworm herbivory, and also tested whether increased susceptibility was associated with lower concentrations of glycoalkaloids. Twelve varieties were originally screened across a range of experimental scales, including laboratory and tunnel experiments and a large‐scale field trial involving over 2000 tubers. 2 In laboratory no‐choice tests, Maris Peer, Marfona and Rooster varieties were significantly more susceptible to wireworm attack, with 63% of tubers showing damage, compared with just 15% of the less susceptible varieties of King Edward, Nadine and Maris Piper. There was also greater tissue consumption and weight gain when wireworms were reared on the most susceptible varieties. 3 In choice tests, wireworms showed a significant preference for those varieties previously identified as being the most susceptible to wireworm herbivory (4.2 holes per tuber) compared with the least susceptible (1.2 holes per tuber). Similar patterns of susceptibility were seen in the field trial, although there was generally more variation in susceptibility. 4 In a tunnel experiment, Marfona and Maris Peer were significantly more susceptibile to wireworm attack (47% of tubers showing damage) compared with Nadine, King Edward and Maris Piper (27% of tubers showing damage). Although Nadine, in particular, had the highest glycoalkaloid concentrations (309.33 mg/kg) and lowest amounts of wireworm herbivory, the relationship between susceptibility and glycoalkaloid concentrations was weak, suggesting that this is unlikely to be the sole mechanism underpinning varietal susceptibility.     

Transitional sublethal and lethal effects of insecticides after dermal exposures to five economic species of wireworms (Coleoptera: Elateridae)

During an insecticide toxicity study involving field-collected dusky wireworm, Agriotes obscurus (L.) (Coleoptera: Elateridae), wireworms exposed dermally to six classes of insecticides exhibited characteristic transitional symptoms of toxicity. These symptoms, collectively termed "morbidity," were categorized as "writhing," "leg and mouthpart movements," or "mouthpart-only body movements." These symptoms could persist for long periods, depending on insecticide and dose, with morbid wireworms ultimately recovering or dying. Additional LC50 and LD50 toxicity studies showed that these stages of morbidity also occurred in four other wireworm species, notably Agriotes sputator (L.), Limonius canus LeConte, Ctenicera pruinina (Horn), and Ctenicera destructor (Brown). In addition, all species exposed dermally to clothianidin moved in significant numbers to the surface of soil in posttreatment holding cups. This movement was not observed when these species were exposed to chlorpyrifos or the control solvent. These findings suggest that toxicity trials involving wireworms should include observations on morbidity, and the duration of trials should continue until symptoms of morbidity cease. The long-term morbidity and potential recovery or death of wireworms exposed to certain insecticides has implications for how laboratory and field studies can be better designed and interpreted in the future.     

WIREWORM POPULATIONS IN RELATION TO CROP PRODUCTION: IV. POPULATION CHANGES DURING A BARE FALLOW

Changes in wireworm populations during a bare fallow were studied on twenty-three fields in 1943 and 1944. In every case populations were much reduced, sometimes to less than 10% of the original level. Wireworms of all sizes were affected, but the larvae less than 5 mm. in length were practically eliminated from the populations.
Evidence is presented that the reduction achieved is greater in fields ploughed in February and March than in fields ploughed in May.
As the populations in the grass 'controls' decline during the same period the entire reduction cannot be attributed to the fallow. The reduction in the bare fallow, however, differs from that in the controls in that it is permanent and continues through the autumn when the counts on the 'controls' are rising.
The practical implications of the results are discussed.     

DNA barcoding identification and genetic diversity of bamboo shoot wireworms (Coleoptera: Elateridae) in South China

Wireworms are recognized as economically important oligophagous pests of Phyllostachys bamboo shoot, which have caused a huge economic loss in South China. Studies of bamboo shoot wireworm control and management were seriously hampered by the lack of reliable identification of larvae. DNA barcoding has been proved to be an important and useful tool in species identification of morphologically cryptic insects. Accurate knowledge of damage-causing species and the phylogenic structure of elaterids will provide insight into their sustainable and effective management. Here, we use interspecific variation in COI and EF1-a as a robust method of identification and consider the intra- and interspecific genetic variation of bamboo wireworms. This approach has revealed that three Melanotus larvae occurred in Phyllostachys forest. The dominance of one species in different sampling location suggests there maybe underlying difference in habit and host plant preference. These data provide a starting point for determining the distribution of wireworm species feeding on bamboo shoots in South China. The inclusions would further aid identification of wireworm species and relationships between certain wireworm species with specific environments.     

Prevalence of Libyostrongylus douglassii in commercially reared ostriches in the highveld region of Zimbabwe

A total of 435 freshly dropped faecal samples were collected from 11 randomly selected ostrich farms during September and November 2002 to determine the prevalence of Libyostrongylus douglassii (ostrich wireworm) in the highveld region of Zimbabwe. Samples, which consisted of 339 samples from breeder birds and 96 samples from pre-slaughter grower birds were screened for nematode eggs using the modified McMaster technique before being individually cultured in an incubator at 28 degrees C. Cultures were examined for the presence of L. douglassii third stage larvae (L3). Using faecal egg counts, eight of 11 farms (72.7%) were positive for L. douglassii in breeders but no eggs were detected in the growers. The faecal culture method detected wireworm larvae in the breeding stock of all farms that were surveyed (100%) and five of the eight farms (62.5%) which had grower birds. Libyostrongylus douglassii was detected in all farms (100%) based on the faecal culture method. Libyostrongylus douglassii was detected for the first time in 7 of 11 farms (64%) surveyed. Data from questionnaires designed to assess farm management practices showed that four out of seven (57.1%) of the ostrich producers were unaware of the importance of wireworms in ostriches. The farms did not have a regular deworming programme for their birds and no faecal samples were sent routinely to the veterinary laboratory for screening of wireworms. Wireworm infections were not taken into consideration by farmers during buying and selling of birds.     

Predation of Thereva nobilitata (Fabricius) (Diptera: Therevidae) on Agriotes obscurus L. (Coleoptera: Elateridae)

Little is known about the natural enemies of wireworms (Coleoptera: Elateridae), but there are frequent anecdotal reports of (usually unnamed) stiletto fly larvae (Diptera: Therevidae) preying on various species. We observed larvae of Thereva nobilitata (Fabricius) feeding on larvae of the dusky wireworm, Agriotes obscurus L., during the summer of 2011, in Agassiz, British Columbia. This finding is of interest as: both the predator and the wireworm are introduced species to this area from Europe; T. nobilitata is uncommon in North America; and this predator has not been associated with any wireworm species previously. We observed that larvae of male and female T. nobilitata will feed on various sizes of A. obscurus larvae, most feeding being carried out by the smallest T. nobilitata larvae. These findings suggest future work should assess the potential for therevid larvae as top‐down regulators of Agriotes larvae under field conditions.     

Wireworms and the sugar-beet crop; field trials and observations

Observations on wireworms in the sugar-beet crop during 1937–40 indicated the importance of alternative food, such as buried turf, weeds, excess seedlings and interdrilled wheat in determining the degree of injury to the crop.
In 1938, out of thirty-six recorded fields growing sugar beet after ploughed-up grass only one suffered severe wireworm damage. Six of these fields were selected for resowing with sugar beet in 1939, and all suffered moderate or severe attacks. Several other fields in their first year from grass showed only slight damage. These observations were supported by an analysis of the Norfolk War Agricultural Committee crop records for 1940.
Seven trials were carried out in 1939 to determine the effect of increasing the seed rate from normal (about 14–16 lb./acre) to 17–25 lb., and of interdrilling with wheat. In three trials, increasing the seed rate gave significantly higher plant populations before singling. It gave significant increases after singling in one out of two trials. However, in the one trial harvested the increase in 'washed beet' was only 7.6% and 'total sugar' 15%. Interdrilling with wheat at 40–70 lb./acre gave significant increases in the plant population before singling in three out of five trials, and after singling in three out of four trials. Increased plant populations both before and after singling were obtained in two other trials with alternating treatments. Increased seed rate and interdrilled wheat together gave greatly improved plant populations before singling in each of two trials and after singling in the one trial where such an observation was possible.     

Effect of temperature on the morbidity and recovery of the Pacific Coast wireworm, Limonius canus, following contact with tefluthrin-treated wheat seeds

Late instars of the Pacific Coast wireworm, Limonius canus (LeConte) (Coleoptera: Elateridae), were exposed to wheat seeds treated with tefluthrin at 10 g active ingredient/100 kg wheat seed, for 2 min, at 10, 15, 20, and 25 °C. All wireworms were moribund within 20 min of first exposure and recovered fully within 7 h. Morbidity induction time (i) decreased considerably as the temperature (t) was increased from 10 to 15 °C, but did not decrease further when the temperature was increased from 15 to 25 °C. For each temperature, induction time decreased as wireworm weight (w) increased, so that . A strong negative temperature coefficient was observed in wireworm recovery, the time required for recovery (r) after exposure decreasing as both temperature (t) and wireworm weight (w) increased, so that r = 1102.47(w)^?0.1848 * e^?0.1012t. The temperature at which wireworms contact insecticides in the field may significantly affect the induction and duration of morbidity, and determine whether wireworms will become moribund before they are repelled by tefluthrin. The ability of wireworms to recover from tefluthrin‐induced morbidity may seriously limit the efficacy of this insecticide in reducing wireworm populations in the field.     

Possible aversion learning in the Pacific Coast wireworm

The effects of carbon dioxide and the induction of morbidity on aversion learning in larvae of the Pacific Coast wireworm Limonius canus LeConte (Coleoptera: Elateridae) are discussed. Wireworms preconditioned by exposing them one or four times to odour of Tefluthrin 20SC and Dividend XLRTA [Syngenta Crop Protection (Canada), Inc., Canada] during the induction of temporary morbidity subsequently contact tefluthrin‐treated wheat seeds in soil bioassays for as long as naïve (i.e. not preconditioned) larvae but are repelled four to five‐fold more frequently by Dividend‐treated seeds in soil bioassays than naïve wireworms, suggesting that wireworms are capable of associating a novel odour (i.e. Dividend) with morbidity but require a minimum of 10–15 min subsequent contact time with treated seeds before being repelled. Wireworms preconditioned by exposure to peppermint odour during the induction of morbidity are not subsequently repelled by peppermint odour in soil bioassays, suggesting that wireworms are either not capable of aversion learning or that the presence of a CO₂ source and/or a suitable host plant may override a negative cue (i.e. peppermint odour). In studies conducted in the absence of soil, a host plant and CO₂ production, wireworms are repelled slightly by droplets of 1.0% but not 0.1% peppermint oil. Previous exposure to peppermint odour or contact with peppermint oil‐treated filter paper during one induction of morbidity does not increase the repellency of wireworms to 1.0% peppermint oil significantly. Repellency to 1.0% peppermint oil is almost eliminated when morbidity is induced five times in the absence of peppermint odour but is restored when peppermint odour is present during preconditioning. These findings suggest that wireworm sensitivity to repellent compounds decreases when repeatedly made moribund, although the results are not sufficient to conclude that wireworms are capable of associative learning.     

Soil bioassay for studying behavioral responses of wireworms (Coleoptera: Elateridae) to insecticide-treated wheat seed

A bioassay for observing wireworm behavior in soil is described. The bioassay permits analysis of orientation, feeding, repellency, and postcontact toxicity behaviors of wireworms in response to insecticide-treated wheat seeds. Wireworm positions were recorded every 5 min for 3 h, and the time required to orient to and contact seeds, and the duration of individual feeding events, was calculated. Both avoidance (before contact with seeds) and repellency (after contact) were quantified. A high proportion of Agriotes obscurus (0.95), Limonius canus (1.00), Ctenicera pruinina (0.80), Melanotus communis/dietrichi (0.80), and Hypolithus sp. (0.70) larvae contacted untreated wheat seeds and began feeding within 120 min when seeds were preincubated for 60 min in soil with 20% moisture. A smaller proportion of A. obscurus contacted seeds if seeds were not incubated in the bioassay before wireworm introduction (0.80) or in soil with 10% moisture (0.65). L. canus larvae required a significantly shorter time (25.3 min) to contact seeds if seeds were incubated for 60 min than if seeds were not incubated before wireworm introduction (43.1 min). Wireworms exposed to untreated seeds and seeds treated with the fungicide Dividend XLRTA fed normally (i.e., sustained feeding for at least 60 min), but a significant proportion of wireworms exposed to seeds treated with Tefluthrin 20 CS (containing the synthetic pyrethroid tefluthrin) fed for 15 min or less and were subsequently repelled. Wireworms exposed to Vitavax Dual (containing the organochlorine lindane) were not repelled after feeding and showed symptoms of illness for up to 28 d before making a full recovery (89%) or dying (11%).