Chapter 8

Host-Fungus Relationships


The intimate relationships that have developed between trichomycetes and their arthropod hosts appear to be long-standing ones in terms of evolutionary time. Evidence for such a statement lies in the diversity to be found within the fungal class, the wide range of arthropod types infested, the worldwide distribution of this kind of symbiotic association, and the number of morphological and physiological specializations that these fungi have had to evolve to insure successful adaptation to their respective host types. In the present chapter we will examine these relationships primarily on the basis of studies of uncultured specimens, whereas in the next chapter (Chapter 9) experimental evidence obtained with axenically cultured fungal species will be emphasized.

Much of the older literature, and some of the newer, refers to the trichomycetes as parasites. If one uses this term in a narrow sense to mean that some demonstrable deleterious effect is produced in one of the partners, then it seems clear from currently available information that the relationship is not parasitic in most species. The fungi more properly should be called commensals, with the commensalism obligate on the part of the fungi but not their hosts. Although the commensalistic relationship may apply to most species of trichomycetes, there are several reports of Smittium morbosum that is lethal to mosquito larvae, and it also has been shown that some instars of mosquitoes can be killed by normally nonlethal Smittium species under conditions involving artificial infestation in the laboratory. More recently a significant pathogenic stage has been found in some Harpellales wherein the fungi grow into the ovaries where they produce cysts that prevent egg development, resulting in sterile females that nevertheless "oviposit" the eggs in new aquatic sites. These pathogenic stages will be described in the section below on Pathogenicity.

Alternatively, there are some data, covered in Chapter 9, that Smittium culisetae may supply nutrients such as B vitamins and sterols that improve the development of nutritionally stressed mosquito larvae. We thus find-perhaps not surprisingly in view of the variety of partnerships involved-that some trichomycetes can deviate from the normal commensalism with their hosts. Other exceptions are likely to be discovered, but it probably will be necessary in such instances to be able to grow both partners axenically and under controlled conditions before any subtle beneficial or deleterious effects by the fungi can be adequately demonstrated.
 

Nutritional Relationships


The trichomycetes spend their entire assimilative stage within the gut and therefore obtain all nourishment from substances present in or passing through the gut lumen. As a consequence they could conceivably deprive the arthropod of some nourishment, but there are no data to indicate that this results in any adverse effect. Since most species of trichomycetes inhabit the hindgut where absorption of food is said to be minimal, the deprivation by these fungi may be insignificant. It is often "healthy" populations of arthropods that appear to be most highly infested, but currently there are no studies comparing the condition of naturally infested populations with uninfested ones. Nor would a good correlation of the data necessarily reveal a cause-effect relationship due to the many uncontrolled factors that undoubtedly influence the development of both partners in their natural condition, not the least of which is the presence of other, nontrichomycetous microorganisms within the gut.

There also are no adequate studies concerned with the effects of these gut fungi on reproduction of their hosts. However, observations by the authors over the years on different kinds of arthropods do not suggest adverse effects on reproductive capacity. In the field one can observe heavily infested crustacean females carrying apparently normal loads of eggs, and in the laboratory some species of millipedes and isopods that reproduce well in captivity appear to produce offspring just as well if infested as not. Species of Harpellales occur only in larval stages of insects, and therefore if any measurable alteration of the reproductive rate attributable to their gut fungi were to be found it would likely be through nutritional consequences prior to reaching adulthood. Trichomycetes that infest adult arthropods have never been found invading reproductive organs or eggs, and if they were to affect reproduction one would also expect them to do so indirectly by nutritional deprivation or through other effects on host metabolism.

The mandibulate arthropods that have trichomycetes in their guts are either herbivorous or omnivorous. The guts of actively feeding arthropods always contain some undigested or undigestible solid matter throughout the gut, and this often accumulates as a bolus in the hindgut where the majority of trichomycete species grow. It is not known whether trichomycetes are capable of digesting some of this debris, or whether they rely on the digestive processes of the host and perhaps on other host metabolites secreted or excreted into the gut cavity that could be essential to their development. Although several species of trichomycetes have now been cultured axenically, this has not helped to resolve the problem, because the isolatable species show no peculiar nutritional requirements; they may, in fact, represent anomalies when one considers that isolation attempts with numerous species of Harpellales and all species of Asellariales and Eccrinales using the same basic isolation media have failed. The role that nutrition plays in the adaptation of these fungi to such an unusual fungal habitat thus remains to be elucidated.
 

Pathogenicity


Arthropod hosts of trichomycetes include several groups of economic or medical importance, and it is of interest to know if their fungal symbionts could affect them adversely. Some millipedes, such as Oxidus gracilis, and sowbugs (Isopoda) are considered to be pests in greenhouses and gardens when their populations are high because of damage they may do to plants during the course of their feeding activities. Of economic significance are the gribbles (Limnoria spp.), small wood-boring isopods that in time can destroy wharf pilings and wooden ships. The importance of mosquitoes as vectors of diseases such as malaria, yellow fever, and dengue fever is well known, and considerable international attention has been given to the blackflies that transmit the filaria of onchocerciasis ("river blindness"), now widespread across central Africa and in some parts of Central and South America. Biting blackflies are notorious pests in many parts of the world, and can also cause considerable loss of livestock (Steelman, 1976).

There are several reports that some species of Smittium can be lethal to their larval mosquito hosts, suggesting that biological control may occur to some extent in natural populations. During the course of laboratory experiments where mosquito larvae were being infested with a species of Smittium, tentatively identified as S. inopinatum (= S. culisetae), Coluzzi (1966) in Italy found that an average of 27% (12-87%) of the larvae of Anopheles gambiae died in 23 infestation studies, whereas larvae of Culex pipiens and Aedes aegypti were not adversely affected. He attributed the death of the larvae to occlusion of the rectal ampulla, which resulted in unsuccessful metamorphosis. More recently, Dubitskii (1978) in Russia reported up to 80% mortality in laboratory colonies of larval Culex pipiens molestus and Aedes aegypti by a species he identified as Smittium culisetae (but probably not that species, based on his trichospore measurements). He surmised that the fungus was introduced into his laboratory cultures via an infested collection of Aedes caspius. The moribund larvae had a black spot visible externally under the abdominal cuticle consisting of dark fungal hyphae, and the larvae were inactive as if paralyzed.

Sweeney (1981a) found a new species, Smittium morbosum, infecting Anopheles hilli in his Australian laboratory in 1978. Significantly, he was able to obtain an axenic isolate of S. morbosum, thus making it available for future experimental study as a system for biological control in the field. Sweeney's culture of S. morbosum is carried in the University of Kansas Culture Collection as isolate AUS-X-1. Sweeney reported 50-90% mortality of larvae in rearing trays, with most of them dying in the 4th-instar. Infection was maintainable in vivo by placing infected larvae in trays with uninfected ones. The original inoculum may have been introduced into the laboratory with a previous field collection of Anopheles annulipes. Infected larvae of A. hilli had a black abdominal spot visible to the unaided eye in the posterior midgut region at the fifth and sixth abdominal segments, and Sweeney attributed this to a possible melanization reaction of the host resulting from restricted penetration by the fungus into the hemocoel. Growth of the fungus established in the anterior hindgut extended forward and penetrated the cells of the midgut epithelium, as well as cells of the Malpighian tubes in some cases. In moribund larvae the fungus seemed to block the hindgut completely, and death also was attributed to the inability of larvae to shed their molts completely. Of special interest is Sweeney's report that hyphal penetration of the midgut epithelium as well as sporulating thalli also occurred in pupae and in adults of both sexes that emerged from trays of infected larvae. This was the first report of Harpellales in pupae or adults of insects (before the discovering of harpellid ovarian cysts in blackflies described below), and such a situation could be a means of transmitting the fungus to new breeding sites in this fungal species. Smittium morbosum is similar but not identical to S. culisetae in trichospore size and shape, and it is possible that the infections Dubitskii and Coluzzi attributed to S. culisetae were in fact caused by S. morbosum or another as yet undescribed species.

The first report of S. morbosum in natural populations of mosquito larvae was by Sato et al. (1989) in Aedes albopictus and Culex pipiens in Japan, followed by the discovery of the fungus in at least five species of mosquitoes belonging to several genera in Argentina (López Lastra, 1990; García et al., 1994). It is evident that S. morbosum can infect a wide range of Culicidae, but the distribution, though widespread, is very disjunct as we currently know it.
 

Ovarian Fungal Cysts


A significant and interesting pathogenic stage of Harpellales was the discovery of ovarian cysts in Simuliidae. Fungal "bodies" were described by Garms (1975) in the ovaries of adult blackflies from Guatemala and Liberia. Nolan (1977) and Yeboah et al. (1984) found several types of cysts in Newfoundland, Canada, blackflies which they believed to be zoosporic "phycomycetes." It was Moss and Descals (1986) in England who demonstrated that germinated cysts found on the surface of oviposited simuliid egg masses in the United Kingdom were the trichomycete Harpella melusinae. Labeyrie et al. (1996) reported H. melusinae and Genistellospora homothallica cysts in the ovaries of New York state blackflies, the species determined by germinating the cysts in water resulting in the production of trichospores (Fig. 8.1, Fig. 8.2, Fig. 8.3). Those two unculturable species have now been identified in Newfoundland by sequences of 18S and 28S rDNA amplified from ungerminated ovarian cysts compared with DNA extracted from morphologically identifiable fungi within larval guts. Moss (1998) has reported germinated cysts attached to Chironomidae eggs whose trichospores resemble Smittium simulii, and there are other patterns of germination indicating that other blackfly gut fungi invade the ovaries as well. These are currently being studied in the authors' laboratory using molecular techniques.

Rizzo and Pang (2005) reported that ovarian cysts of Harpella melusinae found on simuliid egg masses produced one to multiple extensions that developed into zygospores rather than trichospores (normally two per cyst). Because zygospores in H. melusinae develop within the medigut of their simuliid hosts preceded by conjugation of the unbranched thalli, and are believed to be haploid at maturity following meiosis (Moss and Lichtwardt, 1977), this unusual zygospore development from cysts needs to be studied cytologically

There may be only certain times off the year when these fungi grow from the gut into the developing ovaries in the larvae, then through the pupal stage (Yeboah et al., 1984), with cysts forming in the adult ovaries. This is based on limited evidence of cyst frequency in female populations (Yeboah et al., 1984; Labeyrie et al., 1996). In southern United Kingdom Moss and Descals (1986) found cysts in October and November, whereas in New York State Labeyrie et al. found considerable variation in streams sites, with infection rates, based upon relatively few captured females, ranging from 2 to 80%. Yeboah et al. reported infections rates in May at one stream site of more than 50%, but usually lower in other sites and other collection dates. The stimuli or factors that lead the fungus to begin growing into the ovaries is unknown.

This ovarian cyst stage has considerable consequences for both the fungi and their insect hosts. The fungal infection obviously periodically reduces the fitness of simuliid populations in some sites. From the point of view of the symbiotic relationship, it means that the fungi shift from being a presumably benign commensal into a parasite that sterilizes the female. It is possible that this form of parasitism evolved as a requirement for survival of the species. All Harpellales live in the guts of immature aquatic insects, and there is no evidence that fungal propagules in lotic insects can disperse upstream or to other drainage systems on their own. Mosquitoes, bloodworms, ceratopogonids, and other dipterans that breed in still waters have been found with fungi in such sites as tree holes, rock pools, discarded cans, and cemetery vases (e.g., refer to species such as Smittium culisetae, S. culicis, Stachylina grandispora, and Carouxella spp.). Thus, dispersal by flying adults, which "oviposit" the cysts can account for the presence of these harpellids in the headwaters of streams, disjunct lentic habitats, and widespread occurrence of some species in many parts of the world. It is hypothesized that perhaps all species of Harpellales have evolved this or very similar methods of phoretic dispersal by means of adults. If so, ovarial transmission is still to be found in a variety of aquatic insects.

Hindguts of other arthropods are sometimes so packed with trichomycete thalli, most notably millipedes infested with Enterobryus spp., that they would appear to occlude passage of materials through the gut. Thalli in some millipedes may be especially numerous near the pyloric sphincter in the most anterior portion of the hindgut. But there are no reports of death resulting from blockage of the gut in these naturally infested arthropods. It is possible to kill mosquito larvae under some circumstances through overinfestation by cultured Smittium spp., as described in Chapter 9, but conditions that lead to fatalities of this sort are not likely to arise in the natural environment.
 

Ecdysis: Effects on Fungi


All trichomycetes, attached as they are to arthropod cuticles, are subject to periodic separation from their living hosts at the time of molting. The linings of the gut are of ectodermal origin and are shed along with the outer cuticle or exoskeleton. Consequently, the adaptive success of trichomycetes includes means by which they can overcome this periodic expulsion. The intermolt period can be as short as 1 day in hosts such as 1st instars of mosquitoes, or up to many months depending on the kind of arthropod, its developmental stage, and environmental conditions. Some Harpellales are subjected to frequent moltings since they occur in aquatic insect larvae with short developmental cycles. Species of Asellariales and Eccrinales are commonly found in both immature and mature forms of their predominantly millipede and crustacean hosts; in these classes of arthropods molting may continue after sexual maturity has been reached, as the adults continue to grow in size. Reinfestation after molting may be accelerated in certain diplopods and crustaceans that habitually consume a portion of their own exuvia.

Due to the absence of cultured material of relatively few trichomycete species and the fact that fungal growth cannot be observed in the guts without dissection, little is known at present about the rate of development of most trichomycetes in their hosts. Williams and Lichtwardt (1972a) demonstrated that Smittium culisetae could grow and sporulate in 1st instar larvae of Aedes aegypti within 24 hr, which is the duration of that instar under good rearing conditions. Species of most other trichomycete genera appear to develop spores much more slowly. Juvenile forms of arthropods such as isopods and millipedes can be infested with thalli, but the fungi in their guts may not achieve reproductive maturity before being sloughed off with the molt, consequently, sporulating thalli are seen more commonly in mature specimens. In any event, it is clear that rates of development of all species of trichomycetes must be sufficiently in phase with the molting cycles of their hosts for them to have survived.

The molting process has an effect on the reproductive development of many trichomycetes species. In some Harpellales this is seen as a shift from asexual to sexual reproduction. As a consequence, the shed molt of the last larval stage may contain masses of zygospores and no trichospores, and the same situation can be found if these larval forms are dissected just prior to ecdysis. This abrupt morphogenetic shift can be found in some fungal genera (e.g., Legeriomyces, Zygopolaris, Simuliomyces, Glotzia), but not all species of Harpellales respond in the same manner. For instance, thalli of Pennella simulii and Genistellospora homothallica are found producing both zygospores and trichospores concurrently (Williams and Lichtwardt, 1971; Lichtwardt, 1972), and in these cases it is possible that zygospore production is not so much a response to molting stimuli as it is that the last intermolt period is sufficiently long to accommodate zygospore formation. The production of zygospores in the last larval stage may be of special significance in univoltine species of aquatic insects, assuming that zygospores survive longer than trichospores outside the host and can serve as a source of inoculum for the next generation of larvae.

In the Eccrinales, which have no verified sexual state, the response to molting in some species is the production of thick-walled (resistant) primary infestation sporangiospores. For instance, the resistant spores of some species of Arundinula, Astreptonema, Eccrinidus, Eccrinoides, and Taeniella seem to be produced only in response to the molting of their respective freshwater, marine, or terrestrial hosts, but the frequency and abundance of resistant spores in the exuviae can vary with the species of fungus, the degree of maturity of the arthropod, and the season when it molts (Manier, 1969b; Hibbits, 1978). Although ecdysis can be recognized in most of the millipede and crustacean hosts just prior to its occurrence, it is not easy to detect externally when the earlier stages of the molting process have begun and when the fungi presumably are stimulated to start producing resistant spores. Consequently, searching the gut cuticle in fresh exuviae is often the most productive way to find resistant spores. In genera that do not have thick-walled primary infestation spores (e.g., Enterobryus), the production of thin-walled primary infestation spores seems to be less correlated with molting, or not at all.

The clearest evidence for the effect of host development on fungal reproduction is found in the Amoebidiales. The ectocommensal Amoebidium parasiticum produces simple thalli, which convert their entire contents into walled sporangiospores as they mature during the intermolt phase. These sporangiospores serve to establish new thalli on other animals. On molting, injury, or death of the host, developing thalli become converted within minutes or hours into naked amoebae that are released and go through the resistant amoeba-cyst phase of the life cycle. The endocommensal species of Paramoebidium have only the amoeba-cyst phase, and amoebae normally develop and are released only when the host molts or on injury such as caused by dissection. The stimulus for amoeba production in pure cultures of  A. parasiticum has been investigated by Whisler, and his interesting studies are reported in greater detail in the section on amoebagenesis in Chapter 9. Suffice it to say here that Whisler (1966) found he could induce amoebagenesis with an aqueous dyalizate of dried daphnia, and later (1968) determined that amoebagenesis was obtainable in a defined medium so long as the cultures were provided with glucose, amino acids, and relatively high levels of calcium (0.01 M).

One would expect among arthropods to find a common stimulant for amoebagenesis in A. parasiticum, considering the lack of specificity of this species (on Cladocera, Copepoda, Isopoda, Amphipoda, and Insecta), if one assumes (but it has not been demonstrated) that all of the various hosts are capable of inducing the amoeba-cyst phase of the cycle. Whisler (1968) found that crab extract and horse serum, if supplemented with calcium, would induce some amoeba release, and Kuno (1973) used a dialyzed homogenate of mosquito larvae. In the more host-specific endocommensal trichomycetes, nothing is known about the chemical nature of zygospore or resting spore induction. Cultures of Smittium could provide a convenient system to experiment with, but zygospores at present have not been produced in any of the available isolates, and those most used experimentally (S. culisetae, S. culicis. S. simulii) are not known to produce zygospores in vivo either, except for one report in S. culisetae (Williams, 1983b).

Molting hormones could have a direct effect on the fungus, important as they are in metamorphosis, but the molting process is very complex, involving many changes in metabolism, initiation of cellular growth, production of new enzymes, arrestment of feeding, and absorption of proteins and other substances including calcium (in some Crustacea) from the cuticle. Kuno (1973) attempted to get amoebagenesis in A. parasiticum using ecdysterone [as has Lichtwardt (unpublished)], diosgenin, and juvenile hormone, but without success. Hormones acting directly on the fungi may not be the answer in view of Whisler's (1968) amoebagenesis study on A. parasiticum grown axenically on a defined medium, but one cannot necessarily extrapolate from this information and apply it to the endocommensal forms. The natural inducer by which arthropods exert a morphogenetic influence on their gut fungi thus remains an interesting unsolved problem, but not unsolvable provided adequate culture systems become available.
 

Survival and Distribution Mechanisms


The wide geographic distribution of many trichomycete species and the frequency with which they infest arthropods attest to their successful dissemination. Distribution of trichomycetes is determined by their ability to transmit spores from host to host over short or long distances as well as by the motility and migratory activities of the hosts themselves. Most species of terrestrial trichomycetes (Eccrinales and Asellariales) grow in mature, or both mature and immature, stages of their hosts. Many of these terrestrial arthropods (e.g., millipedes, isopods) overwinter in such stages in temperate regions while carrying the fungi in their guts, obviating the need for resistant spores that can withstand the rigors of the external environment over prolonged periods of time. Distribution of trichomycetes among their one or more species of terrestrial hosts appears to be limited only by sufficient proximity of individuals to enable transfer of inoculum, and the range of the populations of arthropods. At times, terrestrial arthropods along with their gut fungi can be carried over great distances by one means or another, as has presumably happened, for example, with their introduction into Hawaii (see Chapter 5).

Marine arthropods with gut fungi (Eccrinales and Asellariales), most of which are intertidal, are comparable to terrestrial ones in many respects. It is the sexually mature stages that are most highly infested and bear the majority of spore-producing thalli. Although individual populations of given arthropod species along coastlines may be disjunct [e.g., fiddler crabs or ghost shrimps (anomurids) that are restricted to coastal mud flats or estuaries], experience has shown that their respective fungi are present in many populations to one degree or another. Overwintering of the fungi, as in terrestrial arthropods, can occur in the guts of the hosts.

The known distribution of the marine Eccrinales and Asellariales, sketchy as collections have been, suggests that some of these fungi may have a worldwide distribution, but they occur in several different species of their particular host types whose geographic ranges are more limited than their gut fungi. This raises the question concerning how these obligate commensals have become disseminated from one ocean to another. Several examples are given here to illustrate the problem. Palavascia sphaeromae has been found in different species of Sphaeroma (Isopoda) collected in the Mediterranean, on the east coast of the United States, and in Japan (Lichtwardt, 1995). In Japan (Hokkaido) P. sphaeromae was also found in another isopod, Tecticeps japonicus, occurring intermixed with an infested population of Exosphaeroma oregonensis. Asellaria ligiae occurs in different species of isopods of the genus Ligia, and is known from the Mediterranean, the east and west coasts of the United States, Puerto Rico, Japan, and in an apparently undescribed freshwater species in Hawaii (see Chapter 4). Enteromyces callianassae has a wider host range; it is known to occur primarily in sluggish, mud-burrowing anomurids (Callianassa and Upogebia species) from Chile, California, and the northwest coast of the United States, but has been found also in true crabs, namely Uca pugilator in France and Hemigrapsus penicillatus in Japan. In the case of E. callianassae, it is tempting to think that dispersal has been via the more active and mobile true crabs, rather than the anomurids.

There is no reason to believe that the distribution of marine trichomycetes has occurred by long-distance dispersal of fungal spores. Rather, it is more reasonable to assume that these gut fungi have dispersed from habitat to habitat over time through migratory activities or accidental displacement of the crustaceans or their molts, and that the fungus has been transmitted from one host species to another while they are living in common habitats. In those trichomycetes (e.g., Asellaria, Palavascia) whose range of hosts consists essentially of several species of one genus, it is conceivable that the fungi originally became established in the more primitive forms, and that the fungi later became distributed as host speciation evolved and the derived species of hosts dispersed.

The most puzzling questions relating to survival and distribution of gut fungi are found in Harpellales and Paramoebidium spp., which live in freshwater larval insects but not in their respective adults. Some species can be found in permanent lakes or ephemeral pools, but the majority live in rapidly flowing streams or rivers. Infested lotic hosts have been found to be common in the headwaters of scattered river systems in the holarctic. In the United States, they are prevalent on both sides of the continental divide in the small mountain streams that initiate major river systems. The larval insects are incapable of migrating any significant distance upstream and, intentionally or not, often drift downstream during the course of their feeding and development (Davies, 1976; Hynes, 1976). Yet, infestation by trichomycetes has been found over a period of years in some of the same upstream sites. Headwater habitats at high altitudes and latitudes may contain no larval hosts during the winter months because of unsatisfactory environments for larval survival (freezing, drying) or, more commonly, the cyclic nature of their development. Maintenance of trichomycetes in the same site year after year can be accounted for either by resistant spores left in the site for renewed infestation, or by recruitment of new inoculum from other localities. In the event of a resident inoculum, one still has to invoke some mechanism of transport over distance to bring about initial establishment of the fungi, and the same situation applies to the many other river habitats where the insects have multivoltine cycles or extended larval development such that overlapping larval stages occur throughout the year.

A few examples will illustrate the degree to which some Harpellales and Paramoebidium spp. can be distributed. Harpella melusinae is one of the most widespread species of trichomycetes, and attaches only to the peritrophic membrane of Simuliidae. In the authors' collecting experience, it can be found in the majority of larval blackfly populations in Europe, the United States, and Japan. Likewise, few streams with stonefly, mayfly, and blackfly larvae do not contain Paramoebidium spp. The successful dissemination of these fungi, both among and within populations of larvae, has been well established. Smittium culicis can be found in many mosquito breeding sites in southern France (Tuzet et al., 1961). On the North Island of New Zealand and especially in eastern Australia, Smittium culisetae, S. simulii, and Stachylina grandispora are common in their respective endemic dipteran larval hosts (Williams and Lichtwardt, 1990). The occurrence of species of Smittium in mosquito populations in other parts of the world may be more sporadic. Nevertheless, when they are found they may occur in rather unexpected places. In Hawaii (Oahu) the senior author isolated S. culisetae from three of three larval sites examined: in discarded beer bottles and cans (containing Aedes albopictus), in a temporary roadside ditch, and in a small rock hole within a forest (both containing A. vexans). In Japan S. culisetae was found in mixed species of mosquito larvae in bamboo and pottery flower vases in a Buddhist cemetery, and S. simulii (normally found in larval blackflies and chironomids) in mixed species of mosquito larvae in a garden watering can (Lichtwardt et al. 1987). Almost all mosquito larva populations in rock holes along stream banks in New South Wales and Queensland, Australia examined proved to be infested with S. culisetae (Lichtwardt and Williams, 1990). In Harpellales, these spotty distributions may be attributable to flying adults carrying the fungi as cysts in their ovaries, but it has yet to be demonstrated that this type of infection occurs in mosquitoes and other aquatic insects, except for blackflies and chironomids (see preceding section on Pathogenicity).