UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of water mite parasitism on Cenocorixa spp. (Heteroptera: Corixidae) Bennett, Andrew M. R. 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1993_fall_bennett_andrew.pdf [ 5.11MB ]
Metadata
JSON: 831-1.0086201.json
JSON-LD: 831-1.0086201-ld.json
RDF/XML (Pretty): 831-1.0086201-rdf.xml
RDF/JSON: 831-1.0086201-rdf.json
Turtle: 831-1.0086201-turtle.txt
N-Triples: 831-1.0086201-rdf-ntriples.txt
Original Record: 831-1.0086201-source.json
Full Text
831-1.0086201-fulltext.txt
Citation
831-1.0086201.ris

Full Text

Effects of Water Mite Parasitism on Cenocorixa spp. (Heteroptera: Corixidae). By  Andrew M. R. Bennett B.Sc. (Hons), U.B.C., 1988  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Zoology) We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA August 1993  © AndrewBenet,193  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  (Signature)  Department of  Zoology  The University of British Columbia Vancouver, Canada  Date^September 28, 1993  DE-6 (2/88)  ii  ABSTRACT The purpose of this study was to understand the mechanisms by which parasitic water mites exclude one species of sympatric water boatmen from low salinity, while a sibling species survives. Attachment of the larval water mite Eylais euryhalina Smith was studied in the laboratory on two species of flight polymorphic water boatmen, Cenocorixa bifida hungerfordi Lansbury and C. expleta (Uhler). After 24 hours of exposure, prevalence and abundance of mites did not differ significantly between host species or host morph (sclerotized versus nonsclerotized). From this, it was concluded that mites recruited to all host types at the same rate. By measuring prevalence and abundance of attached mites only, it was determined that the number of mites initially able to attach also did not differ significantly between hosts. In analyzing the initial location of attachment of E. euryhalina on the four host types, no significant difference was found between species, but a significant difference was discovered between sclerotized and unsclerotized morphs. This effect was evident as a shift of mite attachment from the centre of abdominal segments 2, 3, and 4 on the non-sclerotized hosts, to the thoracico-abdominal membranes (T.A.M.) on the sclerotized hosts. It is speculated that the thickness of the flying hosts' sclerotized integument forces this change in location of mite attachment. A six- to eight-day study of the morphs of each species that are predominant in the field found significant differences in mite prevalence between hosts. Non-flying C. expleta had significantly greater prevalence of mites than flying C. bifida. The number of engorging mites was also significantly greater on non-flying C. expleta. Location of attached and engorging mites followed the same trends as seen in one day experiments. Based on these findings and initial studies, it is argued that it is the sclerotization of C. bifida that causes a reduction in the  iii prevalence of mites over time, rather than a host species effect per se. Because on sclerotized hosts, mites can only engorge on the T.A.M., the number of engorging mites on these hosts is limited to 2 or less, whereas greater number of mites can feed on non-sclerotized hosts. As C. expleta is normally non-flying in the field, whereas C. bifida is predominantly flying, C. bifida has a competitive advantage where mites are present in abundances of greater than 2 mites per host. Field collections of parasitized hosts showed the same patterns of spatial mite attachment as in the laboratory, except that sclerotized hosts often had mites attached directly through the abdominal terga. This must have been the result of mite attachment prior to host sclerotization. Abundances of mites in the field were greater than 2 mites per host in some collections. The predominance of the sclerotized, flying morph of C. bifida appears to allow this species to survive at low salinity where mites abound. C. expleta is excluded from these waters, but its predominantly non-sclerotized, non-flying condition allows better reproduction at moderate to high salinities in the absence of mites. The alternative methods by which these two closely related species of water boatmen have dealt with parasite pressure implicates mite parasitism as a possible impetus in their speciation process.  iv  TABLE OF CONTENTS  Section^  Page  Abstract ^  Table of Contents ^ List of Tables ^  iv vii  List of Figures ^  ix  Acknowledgements ^  xi  Introduction ^  A. Description of Thesis ^  1 1  B. Overview of Water Mite Parasitism ^  11  C. Overview of the Corixidae ^  22  Materials and Methods ^  28  A. Study Site ^  28  B. General Techniques ^  34  1. Corixid Collection ^  34  2. Mite Collection ^  35  3. Infection Experiments ^  39  4. Analysis of Infection Experiments ^  40  5. Long-term Engorgement Experiments ^  42  6. Lake Sampling of Parasitism ^  43  C. Parasitism Parameters ^  44  1. Measures of Quantity of Mites ^  44  2. Measures of Location of Mites ^  46  V 3. Statistical Testing ^ D . Experiments and Field Samples ^  50 50  1. Initial Mite Recruitment and Attachment (1 day) ^ 50 2. Effect of Salinity ^  56  3. Mite Engorgement Experiment ^  60  4. Field Studies ^  63  Results ^  68  A. Initial Mite Recruitment and Attachment (Moderate salinity) ^  68  1. Effect of Host Species ^  68  2. Effect of Host Sclerotization ^  70  B . Initial Mite Recruitment and Attachment (Low and High salinity) ^  77  1. Low Salinity ^  77  2. High Salinity ^  78  C. Long-term Attachment and Engorgement ^  83  1. Attachment ^  83  2. Engorgement ^  84  D . Field Studies ^  89  1. Comparison of Field Parasitism and Laboratory Parasitism ^  89  2. Proportions of C. bifida compared to C. expleta ^ 95 3. Proportions of Flying and Non-flying Cenocorixa spp ^ 97  vi Discussion ^  99  A. Initial Mite Recruitment and Attachment (Moderate Salinity) ^ 1. Effect of Host Species ^ 2. Effect of Host Sclerotization ^  99 99 102  B. Initial Mite Recruitment and Attachment (Low and High Salinity) ^  106  1. Low Salinity ^  106  2. High Salinity ^  108  C. Long-term Attachment and Engorgement ^ 110 1. Attachment ^  110  2. Engorgement ^  111  D. Field Studies ^  115  1. Comparison of Field Parasitism and Laboratory Parasitism ^  115  2. Proportions of C. bifida compared to C. expleta ^ 119 3. Proportions of Flying and Non-flying Cenocorixa spp ^ E. General Discussion ^  120 121  References ^  126  Appendix 1. Mite Life History ^  133  vii List of Tables  Table  Page  Table 1  Mites recorded from Becher's Prairie plotted with respect to salinity ^  13  Table 2  Corixids recorded from Becher's Prairie plotted with respect to salinity ^  23  Table 3  Percentages of flying and non-flying C. bifida in studied lakes from 1962 to 1969 ^  26  Table 4  Dimensions and salinities of lakes studied at Becher's Prairie^1959 - 1969,^1988,^1990,^1991.  33  Table 5  Quantity of parasitism by E. euryhalina on all four host types in Experiment 1 ^  72  Table 6  Quantity of parasitism by E. euryhalina on non-flying C. bifida and non-flying C.^expleta (Experiment 2) ^  74  Quantity of parasitism by E. euryhalina on flying C. bifida and flying C. expleta (Experiment 3) ^  75  Quantity of parasitism by E.^euryhalina on non-flying C. expleta and flying C.^expleta (Experiment 4) ^  76  Quantity of parasitism by E.^euryhalina on all four host types at low salinity (Experiment 5) ^  80  Quantity of parasitism by E.^euryhalina on non-flying C. bifida and non-flying C.^expleta at high salinity^(Experiment 6) ^  82  Mortality of hosts and mites collected from dead hosts over time from engorgement study on flying C. bifida and non-flying C. expleta (Experiment 7) ^  86  Quantity of E. euryhalina attaching over 6 to 8 days on flying C. bifida and non-flying C.^expleta^(Experiment 7) ^  87  Table 7  Table 8  Table 9  Table 10  Table 11  Table 12  viii  Table 13  Table 14  Table 15  Table 16  Table 17  Table 18  Quantity of E. euryhalina engorging over 6 to 8 days on flying C. bifida and non-flying C.^expleta^(Experiment 7) ^  88  Quantity of parasitism by E. euryhalina attaching on flying C. bifida in field collections for the summers of 1990 and 1991...  92  Quantity of parasitism by E. euryhalina attaching on non-flying C.^expleta in field collections for the summers of 1990 and 1991...  93  Quantity of parasitism by E.^euryhalina attaching on flying C. expleta in field collections for the summer of 1991 ^  94  Percent composition of Cenocorixa spp. collected in lakes for summers of 1990 and 1991 ^  96  Percentages of flying and non-flying Cenocorixa spp.^in lakes for the summer of 1991 ^  98  ix  List of Figures Figure  Page  Figure 1  Typical water mite life cycle ^ 15  Figure 2a  Map showing location of Cariboo-Chilcotin region of British Columbia ^  Figure 2b  Map showing location of Becher's Prairie and Kamloops in Cariboo-Chilcotin ^ 29  Figure 3  Map showing the water bodies studied on Becher's Prairie ^  Figure 4  Dorsal plates of the larval mites of the genus Eylais found in British Columbia ^ 38  Figure 5  Recorded areas of attachment of E. euryhalina on Cenocorixa spp ^ 47  Figure 6  Divisions used to plot attachment of E. euryhalina on the terga of Cenocorixa spp ^ 48  Figure 7  Location of attachment on all 4 host types at as shown by percent of host region (Experiment  of E. euryhalina moderate salinity total mites per 1) ^  29  30  73  Figure 8  Location of attachment of E. euryhalina on non-flying C. bifida and non-flying C. expleta as shown by percent of total mites per host region (Experiment 2) ^ 74  Figure 9  Location of attachment of E. euryhalina on flying C. bifida and flying C. expleta as shown by percent of total mites per host region (Experiment 3) ^  75  Figure 10  Location of attachment of E. euryhalina on non-flying C. expleta and flying C. expleta as shown by percent of total mites per host region (Experiment 4) 76  Figure 11  Location of attachment of E. euryhalina on all 4 host types at low salinity as shown shown by percent of total mites per host region (Experiment 5) ^  81  x Figure 12 Location of attachment of E. euryhalina on flying C. bifida and non-flying C. expleta over 6 to 8 days as shown by percent of total mites per host region (Experiment 7)....87 Figure 13 Location of engorgement of E. euryhalina on flying C. bifida and non-flying C. expleta over 6 to 8 days as shown by percent of total mites per host region (Experiment 7)....88  Figure 14  Location of attachment of E. euryhalina on flying C. bifida in field collections ^ 92  Figure 15  Location of attachment of E. euryhalina on non-flying C. expleta in field collections....93  Figure 16  Location of attachment of E. euryhalina on flying C. expleta in field collections ^ 94  Figure 17  Life cycle of E. euryhalina at Becher's Prairie compared to life cycle of Cenocorixa spp. based on field data ^ 134  xi ACKNOWLEDGEMENTS I would like to thank most of all, my parents for their continued support during this thesis, and my grandfather for setting an example for me as a teacher and biologist. A great deal of thanks also goes to my supervisor, Dr. Geoff Scudder for financial support and guidance. Suggestions and reading of this thesis were done by Dr. Martin Adamson and Dr. Murray Isman without whom, the direction of the thesis would not have been acheived. Also, I am endebted to Dr. Bruce Smith who did the original research on the ecology and descriptions of the water mites. Without his work, my thesis would never have been possible. Other thanks goes to Gareth "Oscar Wilde" Williams for company during my first summer at Riske Creek. Also to Louie the cook, and Dr. Locke "Statsboy" Rowe. (Who could mention one of them without thinking of the other?) While at U.B.C., various odd assorted entoweenies kept me giggling such as Karen Needham, Dr. Doug "I always loved Trichoptera" Currie, Dean "Collecting in Edmonton" Mulyk, Charlene Higgins, Jayne Yack, Kathy "Lab Monitor" Craig. People who helped me procrastinate include (but are not limited to) Gregor "18 beers" Reid, Kathy Shimizu, the rest of the Biotechy weenies, the occasionally friendly staff of the Pit Burger Bar and the S.U.B. Arcade, the Columbia Brewing Company, Twevy, Cliffy, Pavy, and the rest of the boys, and of course, Wavy Gravy. (No Art, you're not in here.) Finally, a special thanks to Christina Canil for all her support and attempts to tone down the colour of my t-shirt collection. P.S. It hasn't worked.  1  INTRODUCTION A. Description of Thesis Waterboatmen^(Heteroptera:^Corixidae)^are a common constituent of the insect fauna of ponds, lakes, and slowmoving rivers. The family is notable for its wide-ranging salinity tolerance: from fresh to highly saline waters (Scudder, 1976). Two morphologically similar species of the genus Cenocorixa are abundant in all but the most saline lakes of the British Columbia Interior, occurring in sympatry throughout much of their range. C. bifida hungerfordi Lansbury (hereafter C. bifida) occurs in lakes with conductivities of 20 to 20 000 pS cm -1 , whereas C. expleta (Uhler) normally lives in water with conductivities of 13 000 to 30 000 pS cm -1 . Scudder et al. (1972) found that the exclusion was not caused by C. expleta's inability to osmoregulate at low salinity, in fact Cannings (1978) demonstrated that C. expleta could breed in freshwater in the lab. Nor was the exclusion attributable to differences in life cycle phenologies (Jansson and Scudder, 1974), or because C. bifida outcompeted C. expleta when in sympatry (Reynolds, 1974) (These studies are all reviewed by Scudder, 1983). The most plausible explanation so far forwarded is that of Smith (1977) concerning the effects of parasitic water mites (Acari: Hydracarina) on water boatmen. My thesis  2 is a continuation of this study. Smith (1977), studying the field distributions of C. bifida, C. expleta, and the water mites found that water mites occurred in all lakes below 13 000 pS cm  -1  , but C.  expleta was only found in abundance in lakes above 13 000 pS cm  1  . C. bifida, however, was abundant above and below 13 000  pS cm -1 . It was proposed that mite parasitism had a greater effect on C. expleta compared to C. bifida, resulting in the exclusion of C. expleta in waters below 13 000 pS cm -1 (Smith, 1977). Prevalence of mites in lakes where C. bifida and C. expleta co-occurred with the mites supported this theory. In the pond LE 3 (13 026 pS cm -1 on Sept. 15-20, 1977) parasitism on C. expleta was greater than 40 per cent for 6 samples from July to October. Parasitism of C. bifida in these samples was consistently below 20 per cent. Round-up Lake (10 700 pS cm-1 on Sept. 15-20, 1977) was similar, with more than 25 % parasitism of C. expleta, while less than 15 per cent of C. bifida were parasitized. Parasitism had a direct effect on host fecundity as dissections of field collected parasitized and unparasitized C. bifida showed significant reductions in egg production for the parasitized bugs. With respect to the size of engorging mites on C. expleta and C. bifida, C. expleta was rarely found with fully  3 engorged mites, whereas fully engorged mites were often found on C. bifida. Smith (1977) concluded that the absence of C. expleta with fully engorged mites was because of high mortality of C. expleta prior to full mite engorgement, whereas C. bifida could survive even with fully engorged mites. Laboratory studies by Smith (1977) corroborated the field observations. Larval Eylais euryhalina Smith were used to infect C. bifida from Long Lake (Chilcotin) (8 064 pS cm -1 on Sept. 15-20, 1977) and C. expleta from Barnes Lake (15 197 pS cm -1 ) in fresh, dechlorinated water. When equally exposed to E. euryhalina, over 90 per cent of the C. expleta were parasitized, whereas less than 25 per cent of the C. bifida were parasitized. While there was good evidence that mites had a greater impact on C. expleta than on C. bifida, the actual mechanism behind this effect was not clear. More study was necessary on the corixid-mite interaction in the Chilcotin region to understand how and why C. expleta was more susceptible. The main objective of my study was to determine the exact mechanisms by which water mites limit C. expleta in low salinity lakes. In addition, the life history of the mite Eylais euryhalina Smith was examined in detail. (Data and observations about mite life history are given in Appendix  4  To fulfill my main objective, 5 questions were asked regarding the attachment of the mite Eylais euryhalina Smith (formerly E. infundibulifera # 1) on C. bifida and C. expleta. To answer these questions, hypotheses were formed (see below) and experiments performed (see Materials and Methods). From these experiments, a greater effect of parasitism on one host type was concluded if a.) more mites attached to one host type and/or b.)  a greater susceptible area of mite attachment was  found on one host (explained below: hypothesis 2b).  1.) Does species of host affect a.) quantity of mites initially attaching and b.) where mites initially attach?  la.) Hypothesis (quantity of mites) There are significantly more mites attaching initially to C. expleta than to C. bifida. This would corroborate the findings of Smith (1977) who reported more parasitism on C. expleta compared to C. bifida when equally exposed to four species of mites including E. euryhalina. More mites attaching to C. expleta initially, was predicted to lead to more engorging mites later, causing reduced fecundity and survival.  5 lb.) Hypothesis (location of mites) There is no significant difference between the initial pattern of mite attachment on C. bifida and C. expleta (if host sclerotization is held constant). No difference was expected because the exact position that the mites choose is governed only by the morphology of the air spaces surrounding the hosts' dorsal surfaces, and these regions are similar in both host species (Jansson, 1972).  2.) Does host sclerotization (wing morph) affect a.) quantity of mites initially attaching and b.) where mites initially attach? 2a.) Hypothesis (quantity of mites) There are significantly more mites attaching to unsclerotized (non-flying) hosts than to sclerotized (flying) hosts. The rationale for proposing this hypothesis is that mites are able to pierce the integument of unsclerotized hosts more easily than sclerotized hosts, making prevalence higher on unsclerotized hosts. Since field populations of C. expleta are predominantly non-flying (unsclerotized), whereas C. bifida populations are predominantly flying (sclerotized),  sclerotization could be a factor in the exclusion of C. expleta in areas where mites occur.  6 2b.) Hypothesis (location of mites) Mites on unsclerotized hosts are uniformly spread over the entire dorsum, whereas mites on sclerotized hosts congregate on areas that remain unsclerotized throughout the host's adult life span. This was proposed because mites are predicted to be unable to pierce the hardened terga of sclerotized hosts, therefore, they will only attach on the small, permanently unsclerotized membranes of these hosts. A smaller susceptible area for attachment on sclerotized C. bifida was predicted to result in reduced overall susceptibility to mite parasitism compared to unsclerotized C. expleta. The rationale for this prediction is that crowded mites compete for space and do not extract as much energy from their hosts as a similar number of evenly spaced mites. Mitchell (1968) reported up to 50 % mortality of Arrenurus spp. mites on the damselfly Cercion hieroglyphicum Brauer when mite crowding occurred.  7 3.) Does salinity affect a.) quantity of mites initially attaching and b.) where mites initially attach?  3a. Hypotheses (quantity of mites) Low salinity (less than 300 pS cm') does not significantly alter the quantity of mites initially attaching, on any host type, compared to attachment on the same hosts at moderate salinity. High salinity (above 18 000 pS cm) prevents attachment on all hosts. This question was asked to ensure that the effects of mites are the same throughout the salinity regime from which C. expleta is excluded, but significantly less above the salinity at which E. euryhalina is found (and C. expleta abounds). The rationale for proposing the first hypothesis is that both C. bifida and C. expleta were found to be able to regulate their body fluids at low salinity (Scudder et al., 1972). From this study, both species appear to have no physiological disadvantage at low salinity, so there is no reason to believe that low salinity would affect one species differently than the other with respect to mite parasitism. The hypothesis regarding high salinity is proposed because E. euryhalina is not found breeding in the field at this high salinity, and it may be because of inability to attach to their hosts.  8 3b. Hypotheses (Spatial Attachment) Low salinity (less than 300 pS cm") does not significantly alter the location of mite attachment from what would occur at moderate salinity on all host types. High salinity (above 18 000 pS cm') also does not alter the location of mite attachment.  These hypotheses were proposed for the same reasons as above.  4.) When does the effect of mites occur, or more specifically does the effect occur initially and/or during mite engorgement for a.) quantity of mites and b.) location of  mites?  4a.) Hypotheses (quantity of mites) More mites attach and engorge on the predominant nonsclerotized morph of C. expleta than on the predominant sclerotized morph of C. bifida over a 6 to 8 day period.  In questions 1 and 2, the quantity and location of initial mite attachment was measured on different host types. By comparing the answers to these questions with that of question 4 (long term engorgement), the duration of the effect of mites could be determined. I formed the hypothesis of question 4a because I believed that the lack of  9 sclerotization in non-flying C. expleta would lead to a higher number of mites being able to attach and engorge compared to the flying C. bifida.  4b.) Hypotheses (location of mites) Mites attaching and engorging on unsclerotized C. expleta are uniformly spread over the entire dorsum, whereas mites attaching and engorging on sclerotized C. bifida congregate on areas that remain unsclerotized throughout the host's adult life span. The rationale for proposing this hypothesis is as above (hypothesis 2b). Sclerotization gives C. bifida some protection from mite parasitism whereas C. expleta has no protection because it is predominantly unsclerotized.  5.) Are the data collected from one- and six-day laboratory experiments representative of mite parasitism in populations of field collected corixids for a.) quantity of mites and b.) location of mites?  5a.) Hypotheses (quantity of mites) In field collections, more mites are found on C. expleta than on C. bifida. More mites are found on unsclerotized hosts than sclerotized hosts.  10 This question was asked to verify that the findings in my laboratory studies had some bearing on the exclusion of C. expleta in the field. Overall, I predicted the quantity of mites on field-collected host types to follow the same trends as seen in the one- and six-day laboratory experiments. The hypothesis regarding species of host is based on the study of Smith (1977). He found more mites attaching to C. expleta than C. bifida in several lake samples in which the two host species occur in sympatry with E. euryhalina. I predicted that more mites would be found on unsclerotized hosts in the field because of their greater surface area that is susceptible to attachment (see hypothesis 2a).  5b.) Hypotheses (location of mites) In field collections, there is no difference in the spatial attachment patterns of mites on C. expleta and C. bifida (field collections are similar to laboratory data,  hypothesis lb). Mites found on unsclerotized hosts are uniformly spread over the entire dorsum (field collections are similar to laboratory data, hypothesis 2b). Mites found on sclerotized hosts are also spread over the entire dorsum (field collections differ from laboratory data, hypothesis 2b).  11 The hypothesis concerning species of host was predicted because the interspecific differences are not great enough to cause differences in spatial mite attachment (see hypothesis lb). Field-collected unsclerotized hosts were predicted to have the same mite attachment as in the laboratory because they remain unsclerotized their entire adult life, therefore presenting the same attachment area to mites as the unsclerotized hosts in one and six day experiments. The sclerotized field-collected hosts were predicted to be different from laboratory experiments because in the laboratory, the attachment of mites occurs  after  sclerotization (by design), whereas in the field, attachment may occur prior, during or after sclerotization. Mite attachment prior to sclerotization could occur anywhere on the host's dorsum, causing potential differences in predicted attachment patterns between field and laboratory infected sclerotized hosts.  B. Overview of Water Mite Parasitism Larval water mites form an ecological group known as the Hydracarina (= Hydrachnellae = Hydrachnidia). According to the classification of Prasad and Cook (1972) , they belong to the Subclass Acari, Order Acariformes, and Suborder Parasitengona although the higher classification of the Acari is  12 much debated (O'Connor, 1984). I. M. Smith and Oliver (1986) have reviewed the literature on larval parasitic water mites and their hosts. At Becher's Prairie, 6 species of mites from 2 genera have been recorded parasitizing water boatmen (Smith, 1977) (See Table 1). Of these, only Eylais discreta Koenike (Acari: Eylaidae) and Hydrachna cruenta Muller (Acari: Hydrachnidae) are Old World species. The Eylais species found in North America are described in Smith (1986), and the Hydrachna species are described in Smith (1987). My investigation is the first since the initial description of the ecology of one of these species: Eylais euryhalina Smith. The distribution of water mites at Becher's Prairie extends from the freshest ponds to waters of about 13 000 pS cm' (moderate salinity) (Smith, 1977) although only two species, E. euryhalina and H. barri Smith are present in moderate salinity. E. euryhalina is the only species of mite that occurs in both fresh and moderately saline water, therefore making it the best candidate for a study across a range of salinities. Members of the genus Eylais parasitize long-lived adult aquatic insects such as those of the family Corixidae. They also parasitize other Hemiptera including giant water bugs (Belostomatidae) (Lanciani, 1969) and backswimmers (Notonectidae) (Stout, 1953), as well as aquatic beetles of  13  Mites  Eylais euryhalina  B  B  Eylais lancianii  ?  B  Eylais discreta  B  Hydrachna davidsi  B  B  B  B  B  B  b  B  B  Hydrachna barri Hydrachna cruenta  ?  B  Cenocorixa bifida  B  B  Cenocorixa expleta  ?  R  B  B  B  B  B  B  B  b  R  b  B  B  B  Corixidae  B  .--1 -,-1  .c^a, c._) En^1  0^'0  of^c^c  c^a)^.-4 o^>1^M^0  '4^t-4^CO^C4  Bodies of Water  Increasing Salinity ^ >  Table 1. Mites recorded from Becher's Prairie plotted with respect to salinity (extracted from Smith, 1977). B = breeding in abundance b = breeding, but not in abundance R = recorded, but not necessarily breeding Cenocorixa spp. of corixids shown for reference to salinity. • Round-up Lake was lower in salinity than Barnes Lake before 1979, but has increased in salinity more than the other lakes in the area, and now has no breeding mites and is home to almost exclusively C. expleta.  14 the families Dytiscidae (Aiken, 1985), Gyrinidae and Noteridae (Piatakov, 1916 quoted from Smith and Oliver, 1986), Hydrophilidae and Hydraenidae (Lanciani, 1970b), and Haliplidae (Nielsen and Davids, 1975). The general parasitic water mite life cycle is depicted in Figure 1 (adapted from Harris and Harrison, 1974). In temperate zones, parasitic water mites overwinter as larvae on their hosts. In the spring, the partially engorged larvae continue their engorgement until they reach a critical size at which time they cease engorgement and pupate into a sessile nymphocrysalis (protonymph) on the back of their host. After a brief period (relative to the larval stage), the developed protonymph breaks through the larval cuticle in which it pupated becoming a free-living deutonymph. The immature nymph actively feeds on ostracods and cladocerans (corixid eggs for the genus Hydrachna) for a short time until it once more pupates, this time on a submerged substrate, forming a teleiochrysalis (tritonymph). The fully developed tritonymph breaks through its nymphal cuticle emerging as a dioecious, octopod adult. Nymphs and adults of the genus Eylais can be identified by their bright red colour, relatively large size (up to 16 mm in length), and their habit of trailing the fourth pair of legs behind them when they swim (Lanciani, 1969). Copulation  Teleiochrysalis Egg Nymph  Nymphochrysalis Pre-larva  Larva  Free-living  C:3  Parasitic  16M  Figure 1. Typical Water Mite Life Cycle. The length of arrows approximates the duration of the respective life stages based on the life cycle of E. euryhalina during the spring and summer of 1991. For comparison, the adult stage was approximately 1 month in duration during June and July, 1991.  16 takes place shortly following emergence after which, the males die and the females begin to oviposit. The bright, red eggs are laid in masses on submerged substrates (See Materials and Methods: Section A. 2.). A single female may lay over 10 000 eggs in this manner over a number of weeks (Davids, 1973 for E. discreta). The eggs develop to active prelarvae within their shells and hatch as hexapod larvae after 24-38 days at room temperature (Nielsen and Davids, 1975 on E. infundibulifera Koenike). The positively phototactic larvae swim to the water surface and skate across the top surface in search of an appropriate host. (Hydrachna spp. mites hunt actively for their hosts or cling to the underneath of the water surface.) Hosts coming to the surface for air are mounted, and after finding a suitable attachment site, the larval mites pierce the insect's integument and begin engorgement. Engorgement can take weeks or months depending on the temperature, but most species of Eylais on corixids appear to be bivoltine in temperate regions (Lanciani, 1970a). The summer generation is associated with the host for less than 2 months, whereas the overwintering generation attaches for much longer. Eylais spp. show the largest increases in size during engorgement of any water mite (Lanciani, 1971b).  17 Larvae of the genus Eylais are not truly aquatic, requiring a constant air supply while engorging. They attach only to areas such as on the thoracic and abdominal dorsum covered by the forewings (subelytral air space), the hindwings or underside of the forewings (which are also in the subelytral air space), or occasionally in the air space between the head and prothorax (Davids et a/., 1977). In contrast, many families of mites are truly aquatic in their larval stage (eg. the family Hydrachnidae). Larval Hydrachna spp. can utilize dissolved oxygen in the water allowing attachment to all surfaces of their host including the exterior of the wings, head and legs (Harris and Harrison, 1974 on H. elongata Smith formerly H. cruenta). Attachment to immature hosts is possible for Hydrachna spp., although full engorgement may not occur before host moulting, causing death of the mite. In contrast, some species of mites can attach to an immature host, transfer to the newly moulted adult after host ecdysis, and subsequently begin to engorge (Abro, 1982 for Arrenurus spp. on the damselfly Enallagma cyathigerum Charp.). (Hydrachna spp. and Eylais spp., are incapable of movement once attachment has occurred.) The location of attachment is usually species-specific. Resource partitioning has been demonstrated for the genus Eylais such that species that could compete, parasitize  18 different hosts, different locations on the host, or at different times of the year (Lanciani, 1970a). The exact location of attachment on the dorsum or wing by Eylais spp. has been documented by Nielsen and Davids (1975) for E. discreta on Sigara striata (L.) and Cymatia coleoptrata  (Fab.). They found that on S. striata, E. discreta preferred tergum 3 followed by terga 2 and 4, while on C. coleoptrata, E. discreta preferred tergum 2, -and then 3. Such exact  plotting of mite attachment positions is important in understanding the effects of mites on their hosts, and the co-evolution of sympatric host-parasite systems. The effects of water mites on their hosts was reviewed by Smith (1988) and Lanciani (1983). Following Smith (1988), the levels of effects can be classified as follows: 1. Effects on individuals. 2. Effects on populations. 3. Effects on communities.  1. Effects on individuals. On an individual basis, the effects range from being apparently harmless as for Hydrachna conjecta Koenike on the water boatman Sigara falleni (Fieb.) (Davids, 1973) to being lethal as demonstrated by Lanciani (1975) on the mite-induced reduction in survival of marsh treaders (Heteroptera:  19 Hydrometridae). The causes of mortality have been linked to upset of water balance caused by parasitism (Smith and McIver, 1984c), and rupturing of the integument in cases of superparasitism of mites on damselflies (Mitchell, 1968). Other important effects of parasitic mites on aquatic insects include reduction in the fecundity of females (Davids and Schoots, 1975; Martin, 1975; Smith, 1977), reduction in the rate of nymphal growth (Lanciani and May, 1982), and reductions in male mating success (Forbes, 1991). 2. Effects on Populations. The effects of mortality on individuals can be witnessed at the population level by comparing field samples to a negative binomial distribution (Crofton, 1971). Lanciani and Boyett (1980) demonstrated that on the mosquito Anopheles crucians Wiedemann there was significant mortality of hosts from the mite Arrenurus pseudotenuicollis Wilson when abundance of mites was greater than 11 mites per host. (The negative binomial predicted that there would be more hosts with 11 or more mites than were found in field samples.) Direct observations of populations have drawn the same conclusions. Fernando and Galbraith (1970) reported the absence of the water strider Gerris comatus Drake and Hottes almost 2 months earlier than usual, in years when early collections detected high levels of parasitism by Limnochares  20 aquatica L. Similarly, early mortality because of mites has been inferred by the absence of old-aged hosts with high prevalence of mites, despite a high mite prevalence on younger hosts (McCrae, 1976 on mosquitoes). Mites may also affect other groups of individuals within a population. Mitchell (1967) found higher mite parasitism on male dragonflies compared to females suggesting that mite parasitism can skew the sex ratio of a population. Martin (1975) found that non-flying Sigara falleni had more Eylais spp. mites on them than flying morphs, implying that parasitism could alter the frequencies of morphs. Finally, the location of a population in its habitat can be altered by mite parasitism. Apart from the work of Smith (1977) on the exclusion of C. expleta from low salinity, Wiegert and Mitchell (1973) have shown a similar habitat restriction with respect to temperature. The brine-fly Paracoenia thermalis Viets inhabits thermal pools. It can only optimize its fitness in areas that are not too hot to allow reproduction, but not too cold or stable to allow parasitic water mites to reach high abundances. 3. Effects on Communities. Because parasitic water mites often attach to a wide range of hosts, they can affect the community composition of their hosts through differential effects. Minchella and Scott  21 (1991) review the importance of parasites (including water mites) in determining community structure. Gledhill et a/. (1982) found differential effects of mites on blackflies (Simulium spp.). They determined that one species of host is less affected by mite parasitism than two other species because of a defensive covering on the pupa. Other studies found that the pupae of the mosquito Aedes cinerus Meigen decrease parasitism by Arrenurus spp. mites through violent shaking behaviour, whereas Aedes communis (Degreer) and Aedes punctor (Kirby) do not shake and are more susceptible to mites in laboratory experiments (Smith and McIver, 1984a). A field study, however, demonstrated the complexity of community interactions between these mites and their hosts. In the field, Aedes cinerus is the most parasitized because A. communis and A. punctor develop later in the year, and in so doing, avoid mite parasitism almost entirely (Smith and McIver, 1984b). From these studies, it can be speculated that mite parasitism may cause alterations in the seasonal phenology of its hosts. The differential effects of mites not only affect the parasitized generation, but may also affect the offspring. Decreased or delayed fecundity of parasitized females can cause a smaller and/or later following generation which may affect competitiveness (Martin, 1975 on water boatmen). In  22 cases where food or other resources are limiting, such an effect could be critical to the survival of one host species in a community.  C. Overview of the Corixidae  The corixids of Becher's Prairie are the most conspicuous insects in the lakes. At times, especially near dusk, one sweeping sequence can yield more than 500 corixids, usually constituting more than 90 per cent of the total fauna collected (personal observation in Barnes Lake). In all, 13 species of corixids from 7 genera have been recorded at Becher's Prairie (Smith, 1977) (See Table 2). This study deals exclusively with Cenocorixa bifida and C. expleta, but Hesporocorixa laevigata (Uhler), Cymatia americana Hussey, and Callicorixa audeni Hung. are also  abundant and have been studied with respect to parasitism (Smith, 1977). The taxonomic differences between Cenocorixa bifida hungerfordi and C. expleta have been determined  morphologically (Jansson, 1972) and through acoustic differences^in male^pre-mating^stridulatory patterns (Jansson, 1973). The life cycles of C. bifida and C. expleta are described by Jansson and Scudder (1974) and are typical of corixids in temperate regions. Adults overwinter with the females undergoing ovarian diapause until spring. The males are  23 Bodies of Water  Corixidae Cenocorixa bifida  B  B  Cenocorixa expleta  ?  R  Cymatia americana  B  B  Hesperocorixa laevigata  B  B  Hesperocorixa vulgaris  b  B  B  B  B  B  R  b  B  B  B  B  B  R  R  R  R  B  B  b  R  b  B  R  R  Hesperocorixa atopodonta Hesperocorixa michiganensis  ?  Callicorixa audeni  B  b  b  Arctocorisa sutilis  ?  b  b?  R  Dasycorixa rawsoni Sigara bicoloripennis  R  b  Sigara decoratella  R  B  Sigara penniensis  ?  b R  b  R  R  R  R  R  b?  b?  R  R •••••••■  2 0^(Ti^0 ,4^cci  Bodies of Water  Increasing Salinity ^ Table 2. Corixids recorded from Becher's Prairie plotted with respect to salinity. (Modified after Smith, 1977). B = breeding in numbers b = breeding, but not in numbers R = recorded, but not necessarily breeding N.B. Because the salinities of the lakes have changed since Smith (1977), some of the breeding records may now be in error. I have tried to integrate my findings of 1990 and 1991, with that of Smith (1977).  24 sexually mature by late winter and copulation occurs shortly after ice break-up.^Eggs are laid on submerged rocks and vegetation (Scudder, 1966). Nymphs hatch in late May (depending on temperature) and progress through five instars becoming adults by the middle of June. At the latitudes of Becher's Prairie only two generations are usual. However, further south a partial third generation may occur (eg. C. expleta in the high salinity lake LB 2 near Kamloops). Water boatmen (including Cenocorixa bifida and C. expleta) often exhibit wing muscle polymorphism. Associated with wing muscle development is a hardening and darkening process (Young, 1965a). This process is most evident on the notum and abdominal segments, although the head and forewings (hemelytra) are also affected. The hardness of the corixid notum and tergum is of direct importance to my investigation because these are the susceptible areas for attachment of Eylais spp. mites (See Introduction, Question 2). Wing muscle polymorphism is generally believed to be found in species that have stable habitats for at least part of their life cycle (Young, 1961). Those that live in ephemeral habitats will usually have a flying morph only. The control of wing muscle development seems to depend on the environment in which the newly eclosed (teneral) adult develops (Scudder and Meredith, 1972). Factors that may be involved include  25 temperature, photoperiod, availability of food, and crowding. For Cenocorixa expleta, 1.5 or more days at 15 ° C gives rise to the flying morph, meaning that this morph only occurs in the late summer when the lakes have warmed.^In the field, populations of C. expleta are predominantly non-flying. ^C. bifida, however, is usually predominantly flying. ^Scudder  (1975) proposed that it is the high percentage of C. bifida's population that develops at warm temperatures that makes it mostly flying. He documented the percentages of flying versus non-flying C. bifida for 1962-63 and 1966 to 1969 for 8 lakes of varying salinities. These data and other unpublished records are shown in Table 3. These records may be compared to the data of 1991 (See Results, Table 18). Scudder (1971), following the methods of Young (1965a), classified the development of the wing muscles of Cenocorixa spp. based on the darkening of the mesonotum.^For convenience, I used this method in my study.^The teneral adult (freshly eclosed) is designated as stage 0 (uncoloured notum). Non-flying individuals exhibit up to stage 2 darkening (partial darkening of mesonotum), while flying individuals progress through stages 3 (more area darkening than stage 2) to 6 (completely dark mesonotum). The abdominal segments usually darken in correspondence to the notum and are completely dark by stage 6. Overwintered non-  ^  26 Autumn 1962^Autumn 1963 Lake^Male (n) Female (n)^Male (n) Female (n) F 80^66^79^78 Lye^(50)^(35)^(70)^(40) NF 20^34^21^22 F 79^71^89^83 Barnes^(34)^(24)^(46)^(59) NF 21^29^11^17 Autumn 1966^Spring 1967 Male (n) Female (n)^Male (n) Female (n) ^ F 59 58^51^52 ^ ^ Lye (112) (101)^(126)^(101) ^ NF 41 42^49^48 ^ F 54 59^96^92 ^ Barnes^(78) ^ (101)^(69)^(134) NF 46 41^4^8 Autumn 1967^Autumn 1968 Male (n) Female (n)^Male (n) Female (n) F 64^52^19^8 Lye^(92)^(70)^(59)^(51) NF 36^48^81^92 F 88^87^34^22 Barnes^(25)^(16)^(44)^(45) NF 12^13^66^78 Spring 1969^Autumn 1969 Male (n) Female (n)^Male (n) Female (n) Lye  F 73^61^25^22 ^ (89)^(132)^(71)^(36) NF 27^39^75^78  ^ F 52^42^37 17 ^ Barnes^(23)^(69)^(8) ^ (36) NF 48^58^63 83  Table 3. Percentages of C. bifida flying and non-flying in studied lakes from 1962 to 1969. Extracted from Scudder (1975) and unpublished data. F = Flying NF = Non-flying  27 flying Cenocorixa spp. sometimes have fully darkened abdomens despite mesonotal darkening characteristic of Scudder (1971) stage 2 development.  28  MATERIALS AND METHODS A. Study Site Research was conducted in the Chilcotin region of southcentral British Columbia, Canada (Figures 2 and 3). Locally, the region is known as Becher's Prairie and is situated about 300 km north of Vancouver and 45 km west of Williams Lake. The area is a plateau above the Fraser River at approximately 1 000 m elevation, and is typified by small bunchgrass prairies amid lodgepole pine and Douglas fir forests. Because of the rolling, post-glacial topography and solid bedrock formations, many small, pothole lakes without inflow or outflow are present, with varying salinities dependent on the composition of the underlying strata. In the summers of 1990 and 1991, air and lake temperatures were similar to previous studies (Smith, 1977; Lancaster, 1985) with lake temperatures ranging from 10 to 20  °  C  for May to September, and air temperature fluctuating from 0 to 35 ° C for the same period. Precipitation in the area for 1990 was higher than usual with 450.1 mm, while 1991 was slightly above average (343.2 mm). The lakes in the area have been studied extensively since the 1950's. Chemically, the low salinity lakes are magnesium or sodium bicarbonate-carbonate and sodium bicarbonate-type waters, but at higher salinities they are sodium sulphate and sodium carbonate (Topping and Scudder, 1977). The salinities and sizes of the lakes have been documented by Scudder (1969a, 1969b). Since the salinities can change over time, recordings of the studied lakes in 1990 and  29  Figure 2. Study Site a.) British Columbia b.) Cariboo-Chilcotin Region 1. Becher's Prairie 2. Kamloops (Lake LB 2) Inset of figure la. depicts Cenocorixa expleta (4X life size). Numbers on thin lines of figure lb. indicate highway numbers.  30  Figure 3. The studied water bodies of Becher's Prairie 1. 2. 3. 4. 5. 6. 7. 8. 9.  Box 27 (Pond) Barkley Lake Near Opposite Crescent (Pond) Greer Lake Near Pothole Lake (Pond) Long Lake (Chilcotin) Lake Lye Barnes Lake Round-up Lake  Lakes are listed in ascending salinity from 1 through 9. Barkley Lake and the pond Near Pothole Lake were studied by Smith (1977), but were not studied in depth in this study.^Their relative salinities are assumed to be similar, (or slightly higher) than previous studies (784 to 942 pS cm 1 and 3841 to 4987 pS cm -1 respectively in 1976/77).  Figure 3.  32 1991 were taken and are presented along with dimensional data in Table 4.^Generally, the salinities of all lakes have been gradually increasing to their highest levels on record. ^The biota of the lakes is also well studied. Reynolds (1979) gives a general account of the Crustacean zooplankton of the area and Scudder (1969b) catalogues some of the more common invertebrates (including Corixidae) with respect to salinity. Reynolds and Reynolds (1975) summarize the aquatic angiosperms. Submergent vegetation densely covers most of the fresher lakes, while those above 7 000 pS cm -1 are devoid of such flora. More specific studies include those of Cannings and Scudder (1978) on the Chironomidae; Cannings and Cannings (1987) on the Odonata; Spence (1979) on the Gerridae; and Scudder and Mann (1968) on the Hirudinea. In general, the diversity of the lakes is inversely proportional to their salinity.  Dimensions Lake  Area (h)  Volume^Avg. Depth (1000 m 3 )^(m)  Salinity (p3 cm -1 ) at Given Dates 1959-1969-1959-1969^1988 (Range) (Mean) (August)  Round-up  30.84  787.6^2.6  7 179  Barnes  17.9  348.4^2.0  11^139  Lye  46.52  1283.2^2.8  6 383  Long (Chilcotin)  Greer  Near Opposite Crescent  Box 27  (similar to Barnes)  N/A  (2 8209 000) (3 37020 000) (4^00012 000) N/A .  15.17  156.8^1.0  1^848  6.88  99.2^1.4  835  (1^5252^240) (415^-^973)  4.3  23.0^0.5  40  (26^- 75)  1990^1990 (May) (August)  1991 (June)  1991 (Sept)  18 742  15 224  22 724  N/A  21 304  17 513  13 687  18 464  N/A  19 528  11 214  9 807  12 072  12 427  13 493  N/A  N/A  9 232  13 848  12 215  6 329  3 923  5 681  N/A  5 823  2 519  2 108  3 905  N/A  4 261  39  49  203  128  227  Dimensions and salinities of studied lakes at Becher's Prairie for^1959^-^1969 Table 4. (Averages and^ranges),^August^1988,^and 1990/91.  34 B. GENERAL TECHNIQUES  1. Corixid Collection Corixids for experiments were collected from Barnes Lake (high salinity) to ensure that they were unparasitized. Bugs were collected with an aquatic sweep net using a consistent sweeping technique for each collection. After a sweep was completed, the contents of the net were dispensed on to a white, dissecting tray to allow for easier identification. Adults were separated from nymphs and the adults that were required for an experiment were placed in insulated flasks with lake water and aquatic foliage for transportation to the laboratory. The remaining bugs were returned to the lake except some final instar nymphs that were kept and reared in wading pools at the lake edge to observe the last moult and subsequent sclerotization of teneral adults. Identification of adult corixids was done mainly in the field, although closer scrutiny in the laboratory was necessary for some newly eclosed adult bugs. Extensive handling, especially of unsclerotized morphs, increases mortality. For this reason, handling was minimized. Identification of species was done according to the overall shape, size, and morph. If identification was still equivocal, then the number of hairs on the pala (distal leg segment) of the forelegs was examined as described in Jansson (1972). Verification of morph and the corresponding wing muscle development was done by examining the darkening of the mesonotum  35 according to Scudder (1971). For the pale, non-flying morphs, teneral specimens were used, up to early stage 2 of wing muscle development with only a slight beginning of darkening of the metathorax and abdominal segment 1. For the dark, flying morphs, only stage 6 individuals were used with entirely dark thoraces, and heavy darkening of at least abdominal segments 1 through 5. Non-flying individuals that appeared to have overwintered were not used for experiments, nor were specimens that displayed full wing muscle development, but had yet to fully darken their anterior abdominal segments. Over the summer, the two species' abundances changed, as did the proportions of flying and non-flying morphs. Experiments were carried out as the numbers of each species and morph became large enough for easy collection. Experiments were started no more than 12 hours after collection. Samples of corixids from the lake were taken to check for parasitism levels (see Methods Section D. 4 for dates). These corixids were kept alive at 5°C until examination because preservation in alcohol whitens the mites, causing them to be harder to see on light-coloured hosts. This makes determination of parasitism rates more time-consuming.  2. Mite Collection Eggs of Eylais spp. were collected from Lakes Greer, Long, and Lye, and identified to species after rearing in the laboratory (See below). Oviposition occurred earlier in Lake  36 Greer than in Lake Lye or Long Lake which corresponded with general trends for the development of Cenocorixa spp. in these lakes. In Lake Greer, eggs of the genus Eylais were laid on unanchored strands of Ruppia sp. and Zanichellia sp. (water weeds) that collected on the surface around the lake edges. The eggs were found cemented to the long, tubular, grass-like strands in masses about 1 to 4 cm in length to a maximum diameter of 0.4 cm. The masses were orange and completely encircled the thin strands of vegetation. In contrast, the eggs of E. discreta were much brighter red when freshly laid and were found in masses up to 20 cm in length and 0.5 cm in diameter. In Long Lake and Lye Lake, eggs were found on the submerged stems of Scirpus spp. and Juncus spp. Collection was facilitated by the fact that egg masses were always clumped on the reeds farthest from the shore, with many egg masses layered on top of each other on a relatively small number of reeds (5 to 20). The reeds were cut as far down the stem as possible, and then transported to the laboratory in lake water in insulated flasks. Egg masses were also found on submerged logs, rocks and other aquatic vegetation, but were not as easily located and collected. The egg-laden vegetation was taken to a trailer at the Riske Creek Forestry Station that served as a laboratory. It was then transferred with the original lake water to transparent, plastic stacking dishes (diameter 25 cm by 10 cm depth). A closefitting lid was placed on the stacking dishes and the water  37 level was kept so that the egg masses remained submerged. Through a hole in the lid, a small aeration tube was inserted and kept bubbling very gently throughout the incubation period. A 40 watt lightbulb, illuminated for 16 hours a day, was kept near the hatching eggs. Because larval mites are phototactic, the light was shone from below the stacking dishes so that on hatching, the larvae would not climb out of the dishes. When hatching occurred, the mites could be seen as a red cloud at the bottom of the dish nearest to the light source. For identification purposes, mites were regularly examined under a compound microscope to ensure that only E. euryhalina was used for experiments. The length of the longitudinal furrows on the dorsal plate was used as the criterion for differentiating E. euryhalina from other species of the subgenus Syneylais following Smith (1986) (See Figure 4). The morphologically similar Eylais (Syneylais) lancianii Smith has been recorded in Lake Greer, but was not found in large numbers, and does not occur in Long Lake or Lake Lye where the majority of the mites were taken for experiments. Larval Eylais (Eylais) discreta are much larger than any of the species of the subgenus Syneylais and is easily differentiated from these with the naked eye. To ensure that only live and active mites were used in all experiments, the light source was moved above the container just prior to host innoculation. A pipette was then used to withdraw only the mites that were gathering close to the water surface. This mite-infected water was placed in small drops on a waxed  38  100 pm  C  Figure 4. Dorsal plates of larval mites of the genus Eylais found in British Columbia (adapted from Smith, 1986). a. Eylais (Syneylais) euryhalina b. Eylais (Syneylais) lancianii c. Eylais (Syneylais) peutrilli (not recorded at Becher's Prairie) d. Eylais (Eylais) discreta  39 dish facilitating counting of the mites under a dissecting microscope. The mites were transferred to the experimental water by submerging the dish. 3. Infection Experiments Water for all experiments was gathered from the pond called Box 27 (< 300 pS cm -1 ), Long Lake (10 000 to 15 000 pS cm -1 ) or Barnes Lake (18 000 to 25 000 pS cm - '). Water samples, taken periodically, were bottled and later tested for salinity with a Bach - Simpson TM conductivity meter. All experimental water was held in 10 gallon carbuoys for five or more days to kill any larval water mites that may have been present in the lake samples. In addition, water was strained just prior to the experiment using 44 pm Nitex TM netting. (Larval E. euryhalina are 50 pm wide.) Two litres of strained water were poured into a white plastic 4 litre ice cream bucket for each experiment. The experiments were run at 20 + 2°C for 24 hours with 16 hours of light. A 12 cm by 12 cm piece of mosquito netting was placed in the bottom of the bucket on which the corixids could cling. Once the appropriate number of E. euryhalina had been added to the water, the corixids were added. Care was taken that the various types of bugs were added to the water alternately so that there was no bias for one species, morph or sex of bug being exposed to the parasites for longer than any other. Equal numbers of males and females were used for all species and morphs in each experiment unless this was impossible because of an limited availability of one sex in collection. In these  40 cases, the nearest equal ratio of males and females was used and was never greater than a 60 percent bias towards one sex. After adding all corixids, a transparent lid was placed over the bucket to prevent exit of bugs or entry of other insects. All experiments were run for 24 hours except the long-term experiments (See Methods, Section B.5). At the end of each experiment dead bugs were removed and placed in labelled, waterfilled vials at 5 ° C. These bugs were not included in any analyses because they were not in contact with the mites in the same manner for the full 24 hours. New, strained water was placed in a second bucket with netting in the bottom and the live post-experimental corixids were then moved to these containers by way of a wide-meshed net. Precautions were taken to prevent water from the experiment being transferred to the new water which ensured that only mites that were clinging to the corixids would be left in contact with their hosts past the experimental period. The corixids were then left at 20 ° C for a further 24 hours after which they were removed to 5 ° C and total darkness until analysis. Analysis was always done within 5 days of the conclusion of the experiment. 4. Analysis of Hosts from Infection Experiments Infected corixids were examined with a dissecting microscope for attachment of mites and verification of sex, species, and degree of sclerotization. Those bugs that had died after the 24 hour exposure period, but prior to analysis were examined first.  41 The hemielytra and wings were lifted to allow examination of the wings, thoracic dorsum, and abdominal dorsum. The number and position of mites were recorded for each corixid. Terminology for corixid morphology was taken from the study of Parsons (1970) on Hesperocorixa. In addition to the position of mites on hosts, attachment of the mites was also determined. Those mites that were walking around when the wings were moved, were recorded as unattached. Those that were not moving were examined more closely by lifting the abdomen of the mite to ensure that the mouthparts were inserted. If the abdomen could be lifted, while the cephalothorax remained in contact with the host, then the mite was considered attached. If the entire mite came free when the abdomen was moved, then the mite was considered unattached. In addition, mites that were obviously dessicated Were considered unattached (for all practical purposes) as were ones in which the mouthparts could be seen as unattached. For the 24 hour experiments, if there was any atypical degree of engorgement (i.e. more than a slight swelling), then the bug was discarded because of the suspicion that the attachment of the mite occurred in the field before collection. Additionally, if any E. discreta or Hydrachna spp. were present, then that particular bug was excluded from analyses. In Barnes Lake (where experimental bugs were collected), this occurred only on flying individuals that had flown in from lower salinity lakes and was very rare (less than 0.1 percent). Any time the  42 identification of a mite was at all uncertain, it was removed and examined under a compound microscope. Those corixids that were still alive at the time of analysis were killed prior to analysis. All bugs were then placed in vials of 95 percent alcohol, or for purposes of photography or mite identification, in Koenike's solution (5 parts glycerine, 2 parts glacial acetic acid, 3 parts water). 5. Long-term Engorgement Experiments To ensure that initial attachment of E. euryhalina led to the onset of engorgement, long-term experiments were conducted. Corixids were infected with E. euryhalina as in the 24 hour infection experiments. Immediately after 24 hour exposure, all bugs were taken out to small enclosures kept in Long Lake. It was decided to do all growth experiments in the lake instead of in the laboratory because the constant movement of water was better for corixid survival, and because lake experiments would be more indicative of actual growth in the field. In addition, all bugs were used, rather than only parasitized ones, because examination for parasites caused significant mortality of the corixids. The enclosures were plastic basins of dimensions 30 cm by 35 cm and 12 cm deep.^Lake water could pass into these basins through windows cut in opposite ends of the basins: 44 pm Nitex TM netting over these windows allowed water to enter, but kept mites and additional food out. A close-fitting lid of mosquito netting was kept over the enclosures to prohibit entry  43 or exit of bugs.^Each enclosure was held in a wooden base weighted to the lake bottom and positioned at the lake edge so that the water level of the basin was half full On each day of the experiment, corixids were fed with copepods (Diaptomus spp.) from Barnes Lake. Individuals that had died in the previous 24 hours were removed and taken to the laboratory for analysis. After the allotted experimental period (6 or 8 days), all corixids were taken from the field and analysed immediately in the laboratory. The duration of the experiment was chosen because mortality proved to be too high if done for 10 or more days (based on preliminary experiments in 1990). Also, after 6 to 8 days, a measurable level of engorgement had already occurred. Experimental bugs were killed at two different times so that sample sizes of infected, engorging mites would be large enough to render average sizes of engorgement for both 6 and 8 days post-infection.  6. Lake Sampling for Parasitism Samples of both flying and non-flying C. bifida and C. expleta were collected from Long Lake and Lye Lake (See Methods Section D.4 for dates). As well, collections from Round-up Lake and Barnes Lake were done to ensure that no parasites were present, while Lake Greer and Near Opposite Crescent Pond were sampled to determine if other parasites apart from E. euryhalina were present at lower salinities. All sampling was done in the same manner as described in Methods Section B.1 (General  44 Techniques:^Corixid^Collection)^and^whenever^possible, collections were made at the same location every time. Analysis of parasitism was also similar to the infection experiments, so that lake sample data could be compared to the experiments. C. Parasitism Parameters  1. Measures of quantity of mites. The rates used to quantify differences in the numbers of mites between host types were chosen to correspond with previous studies of mites on water boatmen (Harris and Harrison, 1974; Smith, 1977; Reilly and McCarthy, 1991), and to follow the recommendations of Margolis et al. (1982) regarding usage of terms in parasitological studies. Prevalence is the proportion of hosts in any given population that are parasitized. Abundance is the average number of mites on each host. They were calculated as follows: Prevalence = Bugs Parasitized X 100 Total Bugs Abundance^Total Mites Total Bugs In each experiment, replicates were performed and the average values of prevalence and abundance for the replicates were used for statistical analysis. Prevalence and abundance are measures that include all of the mites found on the hosts, regardless of whether they are unattached, attached, or engorging. These two rates are, therefore, the measures of recruitment of mites to their hosts,  45 but do not take into account the potential effect that a mite may later have on its host. Nevertheless, I deemed it important to measure mite recruitment between host types, because I wanted to know if the cause of C. expleta's exclusion was because it initially attracted more mites than C. bifida. I also wished to measure initial mite attachment on each host type. By excluding mites that were obviously unattached, I calculated prevalence (attached only) and abundance (attached only). These parameters were calculated as follows: Prevalence = Bugs with Attached Mites X 100 (Attached only)^Total Bugs Abundance =^Attached Mites (Attached only)^Total Bugs By analyzing these parameters, I sought to determine whether C. expleta's  exclusion was a result of mites being able to attach  to it more easily than to C. bifida. Finally, one of my experiments allowed the mites to engorge over 6 days (see Material and Methods: Part D.3). For the engorgement data, additional rates were calculated: prevalence (engorging only) and abundance (engorging only) as follows: Prevalence = ^Bugs with Engorging Mites * X 100 (Engorging only) Total Bugs alive after 3 days Abundance = ^Engorging Mites * (Engorging only)^Total Bugs alive after 3 days * Because some bugs died before engorgement could commence (i.e. 3 days), it was not valid to include all hosts in calculations of prevalence (engorging only) and abundance (engorging only). (Inclusion of all hosts in calculations would class the hosts  46 that died prior to 3 days as negative with repect to mite engorgement, despite not knowing if these mites would have engorged or not.) Analysis of prevalence (engorged only) and abundance (engorged only) allowed me to determine whether C. expleta's exclusion was related to the ability of mites to engorge more easily on it than on C. bifida. 2. Measures of Location of Mites Recall from the Introduction (hypothesis 2b), that the quantity of mites is not the only factor involved in determining if mites are having a greater effect on one host type. Only in uncrowded situations will all mites be able to engorge, so that size of a host's susceptible area is important. Comparing the location of mite attachment between host types allowed me to determine if one host type presented a larger area for attachment. The areas of the hosts' body where attachment occurred are shown in Figure 5. ^The suceptible areas were then subdivided into regions (see Figure 6). Numbering of abdominal segments followed Parsons (1970). Both the thoracico-abdominal membranes (T.A.M.) and the anterior edge of each abdominal tergum were apparently used for attachment more often than would have been randomly expected. For this reason, the T.A.M. (both right and left) was considered as a separate area for analysis, as were the anterior edges of each abdominal tergum. I named the anterior edges with respect to the two adjoining terga, thus  47  1 THORACICO-ABDOMINAL MEMBRANE (T. A. M.) TERGA 2 (MIDDLE AND POSTERIOR EDGE) 3 (ANTERIOR EDGE)  4 HIND 5 HIND  LJ  WING (VEIN) WING (MEMBRANE)  AREAS WITH NO RECORDED ATTACHMENT^III  Figure 5. Recorded areas of attachment by E. euryhalina on Cenocorixa spp.  48  NOTUM-ABDOMINAL SEGMENT 1 (NOT-A.S. 1) ABDOMINAL SEGMENT 1 (A.S. 1) ANTERIOR EDGE OF A.S. 2 (A.S. 1-2) ABDOMINAL SEGMENT 2 (A.S. 2)  ANTERIOR EDGE OF A.S. 3 (A.S. 2-3)  ABDOMINAL SEGMENT 3 (A.S. 3) ANTERIOR EDGE OF A.S. 4 (A.S. 3-4) ABDOMINAL SEGMENT 4 (A.S. 4) ANTERIOR EDGE OF A.S. 5 (A.S. 4-5) ABDOMINAL. SEGMENT 5 (A.S. 5) ANTERIOR EDGE OF A.S. 6 (A.S. 5-6) ABDOMINAL SEGMENT 6 (A.S. 6) ANTERIOR EDGE OF A.S. 7 (A.S. 6-7)  Figure 6. Divisions used to plot attachment of E. euryhalina on the terga of Cenocorixa spp.  49 A.S. 1-2 is the anterior edge of abdominal tergum 2, situated just posterior and partially underneath the posterior edge of abdominal tergum 1. To calculate which regions had the greatest attachment, I analyzed each bug separately and then summed all the attachment records in the replicate. I then knew the total number of mites that had attached to each host region for each replicate. To standardize these data, I then converted the numbers of mites in each region to a percentage of total mites in the replicate as follows: Percent of Mites in Area = Mites Attached in Area X 100 Total Mites Attached The results of all replicates were then averaged. To statistically test differences in mite attachment between host treatments, I decided to study attachment to the T.A.M. because it is the only morphologically conservative region in all 4 host types.^(It is constant in terms of size and sclerotization.)^Differences in attachment to the T.A.M. should, therefore, be a good indicator of differences in overall attachment patterns between host treatments. While comparing only one area does not account for differences in all regions, it approximates attachment differences overall because most of the variability in attachment between host types was found in the T.A.M. (See Results). A further X 2 analysis was performed on the 6 to 8 day attachment and engorgement data to determine if attachment of mites was more on the left, centre or right portions of the hosts' body.  50 3. Statistical Testing All statistical tests were done using Systat  TM  . For all data  in percentage form, percentages were changed to proportions and arcsine transformed before analysis. Paired t-tests were used whenever comparing experiments in which two host types were innoculated with mites together (i.e. in the same bucket). Other tests used are as stated in the Results. For graphical and statistical analysis of the location of attachment, unnatached mites were not included.  D. EXPERIMENTS AND FIELD SAMPLES  1. Initial Mite Recruitment and Attachment (1 day) Of the 5 questions asked in the Introduction, the first 3 pertained to the initial stages of the mite-corixid interaction. Questions 1 and 2 were answered by Experiment 1 (moderate salinity) while question 3, regarding the role of salinity in the exclusion of C. expleta, was answered by a similar experiment done at low salinity (Experiment 5), and another test at high salinity (Experiment 6). In Experiment 1, larval E. euryhalina were offerered equal numbers of each of the four host types in moderate salinity water. It was then possible to study both the effects of host species (Question 1) and host sclerotization (Question 2) on initial mite attachment. By using lake water from Long Lake (10 000 to 15 000 pS cm -4 ), I was able to test the attachment of mites at the highest salinity from which C. expleta is excluded.  51 This could then be compared with a similar experiment at low salinity to ensure that the effects of mites are the same throughout the regime of C. expleta's exclusion (Question 3). Recall from the Introduction:  Question 1.) Does species of host affect a.) quantity of initial mite attachment and b.) location of initial mite attachment?  Question 2.) Does host sclerotization (wing morph) affect a.) quantity of initial mite attachment and b.) location of initial mite attachment? The design of Experiment 1 was as follows: Experiment 1: (2 replicates) Corixids^ 30 30 30 30  C. C. C. C.  bifida teneral non-flying bifida flying expleta teneral non-flying expleta flying  Mites^Salinity 600 mites  Moderate (10 000 to 15 000 pS cm ') -  From Experiment 1, six sets of data were collected to answer the various parts of question 1 and 2. The six questions were as follows: la. Does host species affect quantity of initial mite recruitment? Does host species affect quantity of initial mite attachment? lb. Does host species affect location of initial mite attachment?  52 2a.^Does^host  sclerotization  affect  quantity  of  initial  mite  sclerotization  affect  quantity  of  initial  mite  sclerotization  affect  location of initial mite  recruitment? Does^host attachment? 2b.^Does^host attachment? Note that questions la and 2a have two parts, firstly comparing mite recruitment and then mite attachment. (For simplicity, the hypotheses and questions in the Introduction were stated in terms of attachment only.) For each of the questions above, a data set was collected and compared to the expected results (based on the hypotheses in the Introduction). la. Does host species affect quantity of initial mite recruitment? Data collected: Prevalence and abundance of mites were compared between C. bifida and C. expleta (regardless of host sclerotization). Expected results: Prevalence and abundance of E. euryhalina on C. expleta is greater than on C. bifida. Analysis: Greater initial recruitment of mites to C. expleta is at least partially responsible for C. expleta's exclusion from low salinity.  53 Does host species affect quantity of initial mite attachment? Data collected: Prevalence (attached only) and abundance (attached only) were compared between C. bifida and C. expleta (regardless of host sclerotization). Expected results: Prevalence (attached only) and abundance (attached only) of E. euryhalina on C. expleta is greater than on C. bifida. Analysis: Greater initial attachment of mites on C. expleta is at least partially responsible for C. expleta's exclusion from low salinity.  Question lb: Does host species affect location of initial mite attachment? Data collected: Percentage of total mites collected that attach on the T.A.M. was compared between C. bifida and C. expleta (regardless of host sclerotization). Expected result: There is no significant difference in the percentages of attached mites on the T.A.M. between C. expleta and C. bifida. Analysis: Since there was no difference in the location of mite attachment between host species, this factor is not important in the exclusion of C. expleta from low salinity.  54 Question 2a: Does host sclerotization affect quantity of initial mite recruitment? Data collected: Prevalence and abundance were compared between non-flying (non-sclerotized) and flying (sclerotized) hosts (regardless of host species). Expected result: Prevalence and abundance of mites is greater on non-flying hosts than flying hosts. Analysis: Lack of sclerotization causes higher recruitment of mites compared to sclerotized hosts. Since C. expleta is predominantly unsclerotized in the field, this condition is important in C. expleta's exclusion from low salinity.  Does host sclerotization affect quantity of initial mite  attachment? Data collected: Prevalence (attached only) and abundance (attached only) were compared between non-flying hosts and flying hosts (regardless of host species). Expected results: As for the effect of sclerotization on initial mite recruitment. Analysis: As above.  55 Question 2b: Does host sclerotization affect location of mite attachment? Data collected: Percentage of total mites collected that attach on the T.A.M was compared between non-flying hosts and flying hosts (regardless of host species). Expected results: On flying (sclerotized) hosts, all mites attach to the T.A.M., whereas on non-flying (unsclerotized) hosts, attachment is possible over the entire area of the host dorsum. Analysis: Sclerotization of the host reduces the surface area available for mite attachment. Since C. bifida is predominantly sclerotized in the field, it is less susceptible to mite parasitism than C. expleta which is mainly unsclerotized.  To replicate the results of Experiment 1, three further experiments were run using only 2 of the potential hosts. While it would have been ideal to perform Experiments 2, 3, and 4 with a total of 120 corixids and 600 mites per replicate (as in Experiment 1), the numbers of E. euryhalina hatching at any one time constrained this. ^Each replicate was accordingly halved in size with 60 bugs total and only 300 mites. Experiments 2 and 3 held host sclerotization constant, and were thus only concerned with testing for the effect of host species on quantity and location of mites (Question 1). The  56 designs were as follows: Experiment 2: (2 replicates) 30 C. bifida teneral non-flying 30 C. expleta teneral non-flying  300 mites  Moderate (10 000 to 15 000 pS cm ') -  Experiment 3: (2 replicates) 30 C. bifida flying 30 C. expleta flying  300 mites^Moderate  Experiment 4 held host species constant, and was only concerned with testing for the effect of host sclerotization (Question 2). The design was as follows: Experiment 4: (2 replicates) 30 C. expleta teneral non-flying^300 mites^Moderate 30 C. expleta flying The data collected from these experiments were tested in exactly the same manner against the hypotheses of question 1 and 2.  2. Effect of Salinity Question 3 of the Introduction dealt with the effect of salinity on the exclusion of C. expleta. 3.) Does salinity affect a.) quantity of mites initially attaching and b.) where mites initially attach? The first part of hypothesis 3a. dealt exclusively with low salinity and this factor was studied in Experiment 5. The experiment was the same as Experiment 1 except for salinity.  57 Experiment 5: (2 replicates)  Corixids 30 30 30 30  C. C. C. C.  ^  ^ Low (< 300 pS cm -1 ) expleta teneral non-flying^ expleta flying  bifida teneral non-flying bifida flying  ^  Mites^Salinity 600 mites  The following data were collected to answer the various parts of question 3 at low salinity. Question 3a (low salinity). ^Does low salinity affect quantity of initial mite recruitment? Data collected: Prevalence and abundance were compared between host types and between salinities (for similar host types). Expected Result: Prevalence and abundance show the same differences between host types at moderate and low salinity. From hypothesis la, C. expleta will have more mites recruiting to it than C. bifida at  both moderate and low salinities. From hypothesis 2a, nonflying hosts will have more mites recruiting to them than flying hosts at both salinities. Analysis: Low salinity does not affect the quantity of mites recruiting to different hosts compared to moderate salinity.  58 Does low salinity affect quantity of initial mite attachment? Data collected: Prevalence (attached only) and abundance (attached only) were compared between hosts and between moderate and low salinities. Expected result: As above for initial mite recruitment. Analysis: Low salinity does not affect the quantity of mites attaching to different hosts compared to moderate salinity. Question 3b (low salinity). Does low salinity affect location of initial mite attachment? Data collected: Percentage of total mites attaching on the T.A.M. was compared between host species and between moderate and low salinity. Expected Result: Percentage of total mites attaching on the T.A.M. shows the same^differences between host types at moderate and low salinity.^From hypothesis lb, there is no difference in location of mite attachment between C. bifida and C. expleta at both moderate and low salinity. From hypothesis 2b, significantly more mites attach to the T.A.M. of the flying hosts than the non-flying at both salinities. Analysis: Low salinity does not affect the location of mite attachment, with respect to moderate salinity.  59 High salinity: The same questions were asked of high salinity, but because the expected result from hypothesis 3 was that no mites would attach, this experiment was done with only 2 hosts and half the number of mites.^(It was much harder to set up an experiment with 4 hosts.)^The results were then compared to the similar experiment performed at moderate salinity (Experiment 2). The design was as follows: Experiment 6: (2 replicates) Corixids^  Mites^Salinity  30 C. bifida non-flying^300 mites^High 30 C. expleta non-flying^ 18 000 to 25 000 pS cm -1 )  Question 3a (high salinity). ^Does high salinity affect quantity of initial mite recruitment? Expected Result: Prevalence and abundance are significantly lower (approaching 0 %) at high salinity compared to moderate. Analysis: High salinity decreases the quantity of initial mite recruitment (compared to moderate salinity).  60 Does high salinity affect quantity of initial mite attachment? Expected Result: Prevalence (attached only) and abundance (attached only) are significantly lower (approaching 0 %) at high salinity compared to moderate. Analysis: High salinity decreases the quantity of initial mite attachment (compared to moderate salinity). Question 3b (high salinity).^Does high salinity affect location of initial mite attachment? Since it is predicted that prevalence of mites at high salinity approaches 0 %, a comparison of location of mite attachment is not necesssary.  3. Mite Engorgement Study (6 to 8 days) Question 4 of the Introduction was concerned with mite engorgement. 4.) When does the effect of mites occur, or more specifically does the effect occur initially and/or during mite engorgement for a.) quantity of mites and b.) location of mites? Experiments 1 through 4 dealt with the initial effect of mites on different host types at moderate salinity. Experiment 7 studied the effects of mite engorgement at moderate salinity over 6 to 8 days. While it would have been ideal to do Experiment 7 with all four host types, time constraints did not  61 allow this.^In addition, teneral, non-flying C. bifida harden very quickly creating problems in the analysis of engorgement of mites on flying versus non-flying morphs. I chose, therefore, to use flying C. bifida, and non-flying C. expleta, because these two morphs were most abundant in their natural populations and consequently, most important to study for the effects of mite parasitism on the exclusion of C. expleta. The design was as follows: Experiment 7: (4 replicates) Corixids 60 C. bifida flying 60 C. expleta non-flying  Mites^Salinity 600 mites^Moderate (10 000 to 15 000 pS cm')  Two of the replicates were kept in Long Lake for 6 days after the initial 24 hours, while the other two replicates were kept for 8 days so that size of engorged mites could be compared at these times. The following data were collected to answer the various parts of question 4. Question 4a: Does host affect the quantity of mite attachment over 6 to 8 days? Data collected: Prevalence (attached only) and abundance (attached only) of mites were compared between flying (sclerotized) C. bifida and non-flying (non-sclerotized) C. expleta.  62 Expected results: More mites are able to attach on non-flying C. expleta than on flying C. bifida over 6 to 8 days. Analysis: The predominant morph of C. expleta is more susceptible to mite attachment than the predominant morph of C. bifida over 6 to 8 days. Does host affect quantity of mite engorgement over 6 to 8 days? Data collected: Prevalence (engorged only) and abundance (engorged only) were compared between flying C. bifida and non-flying C. expleta. Expected results: As for the effect of host on quantity of mite attachment. Analysis: As above. Question 4b: Does host affect location of mite attachment over 6 to 8 days? Data collected: Percentage of total mites collected that attach on the T.A.M. was compared between flying C. bifida and non-flying C. expleta over 6 to 8 days. Expected results: On flying C. bifida all mites attach on the T.A.M., whereas on non-flying C. expleta, mites attach over the entire dorsum.  63 Analysis: Sclerotization of the host reduces the surface area available for mite attachment. Since C. bifida is predominantly sclerotized in the field, it is less susceptible to mite parasitism over 6 to 8 days than C. expleta which is mainly unsclerotized. Does host affect location of mite engorgement over 6 to 8 days? Data collected: Percentage of total mites collected that engorge on the T.A.M. was compared between flying C. bifida and non-flying C. expleta. Expected results: As for the effect of host on the location of mite attachment. Analysis: As above.  4. Field Studies. Corixids were sampled throughout the summer at lakes of varying salinity. Lakes to be sampled were chosen based on the parasitological data of Smith (1977) to give an overview of the mite-corixid interaction over a wide salinity range. Corixids were collected for 3 reasons: 1.) To ensure that laboratory experiments were representative of field infections with respect to the quantity and location of E. euryhalina on its hosts (question 5).  64 2.) To determine the relative proportions of C. bifida and C. expleta in lakes (demonstrating that mites are excluding C. expleta where they are present).  3.) To determine the relative proportions of flying and nonflying Cenocorixa spp. (demonstrating that the sclerotization of the host is related to the presence of mites).  Samples were made in the following lakes and on the following dates: Water body^Date^ Salinity Low moderate Near Opposite Crescent Pond September 14, 1991^4 261 pS cm - ' Moderate Long Lake^June 2, 1991^13 848 pS cm -1 July 17, 1991^11 220 to 13848 pS cm -1* September 13, 1991^12 215 pS cm -1 Moderate Lake Lye^October 21, 1990^12 072 pS cm -1 June 2, 1991^11 788 to 13 493 pS cm -1* August 17, 1991^11 788 pS cm -1 September 13, 1991^13 493 pS cm -1 High Barnes Lake^August 14, 1991^18 464 to 19 528 pS cm -1* Round-up Lake August 14, 1991 ^18 820 to 21 304 pS cm -1* *^  Water samples were not taken on these days of corixid  sampling, so the ranges found in that year are given instead.  65 1.) Comparison of field parasitism and laboratory parasitism.  Question 5.) Are the data collected from one- and six-day laboratory experiments representative of mite parasitism in populations of field collected corixids for a.) quantity of mites and b.) location of mites? Of all the field samples, only July 17 (Long Lake), October 21 (Lake Lye), August 17 (Lake Lye), and September 13 (Lake Lye), yielded enough parasitological data for analysis. Smith (1977) gives a much more thorough account of the overall parasitism rates for C. bifida and C. expleta in a number of lakes. The following data were collected to answer the various parts of question 5. Question 5a: Are laboratory experiments representative of the quantity of mites attaching on field collected corixids?  Data collected: Prevalence (attached only) and abundance (attached only) were compared between field collected host types and then compared to similar values for initial (1 day) and long-term (6 day) laboratory experiments. Expected results: Relative quantities of mites attaching on hosts are the same in the field and the laboratory. Analysis: One- and 6-day mite attachment studies in the laboratory are  66 representative of mite attachment occurring in the field with respect to quantities of mites. Question 5b: Are laboratory experiments representative of the location of mites attaching on field collected corixids? Data collected: Percentage of total mites attached on the T.A.M. was compared between field collected host types and then compared to similar values for initial (1-day) and long-term (6-day) laboratory experiments. Expected results: In field collections, location of mite attachment is the same as in laboratory experiments on non-sclerotized C. bifida and non-sclerotized C. expleta (mites attach all over the dorsum). Location of mite attachment on sclerotized C. bifida and C. expleta differs between field collections and laboratory experiments.^In the laboratory experiments, these hosts are predicted to have attachment only on the T.A.M.^In the field, they are predicted to have attachment all over the dorsum. Analysis: Location of mite attachment on the non-sclerotized C. bifida and C. expleta is the same in the field and laboratory because these hosts offer an entire unsclerotized dorsum to the mites whether they are in a laboratory experiment or collected from the field. The sclerotized field-collected hosts exhibit different mite attachment from laboratory experiments because in the laboratory, the attachment of mites occurs after  67 sclerotization (by design) causing all mites to attach on the T.A.M.^In the field, however, attachment may occur prior, during or after sclerotization. ^Mite attachment prior to sclerotization could occur anywhere on the host's dorsum causing the difference in predicted attachment patterns between field and laboratory infected sclerotized hosts. Laboratory experiments are not always representative of location of mite attachment in the field.  2.) Proportions of C. bifida compared to C. expleta. Since the salinities of the lakes have changed since the study of Smith (1977), the relationship between salinity, mites, and corixids was re-examined. This was done by noting the salinities at which mites are present (See Appendix 1), and the relative percentages of C. bifida and C. expleta in the studied lakes (see Results). C. bifida should be predominant where mites are present, whereas C. expleta should predominate above the salinity at which mites exist.  3.) Proportions of flying and non-flying Cenocorixa spp. Since the proportions of flying and non-flying morphs of Cenocorixa spp. change over time (Scudder, 1975), it was necessary to determine what forms were predominant in 1990 and 1991 (See Results). Based on previous collections, it was predicted that C. bifida should be predominantly flying, whereas C. expleta should be predominantly non-flying.  68 RESULTS A. Initial Mite Recruitment and Attachment (Moderate Salinity). 1. Effect of Host Species. Question la. Does host species affect quantity of initial mite recruitment? Data collected: In Experiment 1 (Table 5) with all four host types, prevalence of E. euryhalina was not significantly different between C. bifida and C. expleta (Two-way ANOVA, grouped by host species: F = 0.200, P = 0.678). In Experiment 2 with non-flying hosts only (Table 6) there was also no significant difference, t = -1.230, (P = 0.435), and neither was there in Experiment 3 with flying hosts only (Table 7) (t = -1.849, P = 0.316). Abundance was also not significantly different between host species in Experiments 1, 2, and 3. Summary: In 24 hour laboratory experiments, host species did not affect quantity of mite recruitment. Does host species affect quantity of initial mite attachment? Data collected: For Experiment 1 (Table 5), prevalence (attached only) of E. euryhalina was not significantly different between C. bifida and C. expleta (Two-way ANOVA, grouped by host species: F = 0.001, P = 0.976). Experiment 2 (non-flying hosts only) followed the same trend (t = -1.760, P = 0.329) (Table 6), as did experiment 3 with flying hosts only (t = 0.909, P = 0.530) (Table 7).  69 Abundance (attached only) followed the same trend. Summary: In 24 hour laboratory experiments, host species did not affect the quantity of mites attachment.  Question lb: Does host species affect location of initial mite attachment? Data collected: In experiment 1, the percentage of total mites that attached on the T.A.M. was not significantly different between C. expleta and C. bifida (Two-way ANOVA, grouped by host species: F = 0.174, P = 0.698) (see Figure 7). Neither was Experiment 3 for flying hosts only (t = -5.798, P =0.109) (see Figure 9). Experiment 2, however, had a significantly higher percentage of the collected mites attached on the T.A.M. of non-flying C. bifida compared to non-flying C. expleta (t = 145.4, P = 0.004) (see Figure 8). An overview of the host types shows that flying C. bifida and flying C. expleta had a consistently high proportion of mites initially attaching on the T.A.M., whereas C. expleta non-flying had a consistently low proportion. C. bifida non-flying had attachment proportions on the T.A.M. that were variable and intermediate between the flying hosts and nonflying C. expleta. This variance in the attachment patterns of mites on non-flying C. bifida was consistent throughout all experiments. Summary:  70 On the flying hosts, species of host did not appear to affect the location of attachment of mites (Experiments 1 and 3). On the non-flying hosts, the variance in the attachment patterns of mites on non-flying C. bifida precluded a definite answer of this question (Experiments 1 and 2).  2. Effect of Host Scierotization. Question 2a: Does host sclerotization affect quantity of initial mite recruitment? Data collected: In Experiment 1 (Table 5), prevalence of E. euryhalina was not significantly different between non-flying (non-sclerotized) and flying (sclerotized) hosts (Two-way ANOVA, grouped by host sclerotization, F = 0.010, P = 0.925). The same was true for Experiment 4 (Table 8) with only C. expleta (t = 0.326, P = 0.799). Abundance was also not significantly different between non-flying and flying hosts in Experiments 1 and 4. Summary: In 24 hour laboratory experiments, host sclerotization did not affect quantity of mite recruitment. Does host sclerotization affect quantity of initial mite attachment? Data collected: In Experiment 1 (Table 5), prevalence (attached only) of E. euryhalina was not significantly different between non-flying and flying hosts (Two-way ANOVA, grouped by host sclerotization,  71 F = 0.617, P = 0.476). The same trend was found in Experiment 4 with only C. expleta (t = 1.049, P = 0.485) (Table 8). Abundance (attached only) was also not significantly different between host species in Experiments 1 and 4. Summary: In 24 hour laboratory experiments, host sclerotization did not affect quantity of mite attachment.  Question 2b: Does host sclerotization affect location of initial mite attachment? Data collected: In experiment 1, the percentage of total mites that attached on the T.A.M. was significantly different between non-flying and flying hosts (Two-way ANOVA, grouped by host sclerotization: F = 34.001, P = 0.004) (see Figure 7). So was Experiment 4 with only C. expleta (t = - 57.079, P = 0.011) (see Figure 10). Flying hosts had a high percentage of mites attached on the T.A.M., whereas non-flying hosts had a low percentage, with most mites attaching on the centres (laterally) of abdominal terga 2, 2-3, and 3. Summary: Host sclerotization affected the location of initial mite attachment. Mites on sclerotized hosts attached almost exclusively on the T.A.M., whereas mites on unsclerotized hosts were all over the dorsum.  72  ^ HOST PARASITISM ^ ^ RATES C. bifida C. bifida^C. expleta ^C. expleta ^ ^ ^ flying (AVG. OF 2 REPS) non-flying flying^non flying -  PREVALENCE ( + /-s . E . ) PREVALENCE (ATTACHED ONLY)^(+/-S.E.) ABUNDANCE ( + /-S . e . ) ABUNDANCE (ATTACHED ONLY) (+/-s .E. )  66.8  (+/ - 9.8)  34.7  (+/ - 3.43)  42.1 (+/ 0.0)  80.0  62.3  (+1-9.8)  19.5  (+1-0.95)  34.2 (+/-2.8)  55.4 (+/-1.2)  3.31  (+1-0.59)  0.64  (+/-0.03)  1.00  (+/-0.15)  3.90 (+1-0.15)  2.37  (+/-0.62)  0.35  (+/ - 0.01)  0. 63  (+1-0.09)  1.70 (+/-0.18)  -  (+/-2.8)  Table 5. Experiment 1. Parasitism rates for quantity of mites associated with all 4 host types over 24 hours at moderate salinity (10 000 to 15 000 pS cm -1 ).^All parasitism rates show no significant difference between hosts. ^For statistical analyses, see text.  -  73 T.A.M. NDIUM NOTUM,AS  ,  AS 1 AS 1-2  3  ,  AS 2 AS 2-3  AS 3 AS 3-4 AS 4 AS 4-5 AS 5. Wing (vein)  Q CD Lil  Wing (memb.)  C. bifida  non-flying ^  Mites 72,21 ^ Bugs 18,28  C. bifida  flying Mites Bugs  9,9 24,27  CO^T.A.M. C) MDTUM NDTW-A.S . 1 AS 1 AS 1-2 AS 2 AS 2-3 AS 3 AS 3-4 AS 4 AS 4-5 AS 5. Wing (vein) Wing (memb.)  C. expleta  non-flying  ^ 17,7 Mites ^ Bugs 19,19  C. expleta  flying Mites Bugs  54,21 25,17  0^20^40^60^80^100^20^40^60^80^100 AVG MITES ATTACHED PER REGION (PERCENT OF TOTAL)  Figure 7. Experiment 1. Location of mite attachment on all host types at moderate salinity (10 000 to 15 000 pS cm'), plotted as a percent of total mites collected on each host type. Numbers of mites in boxes are attached mites only (per replicate) associated with number of bugs directly below. Error bars are standard error. LOCATION OF ATTACHMENT as measured by percent of total mites on T.A.M. Two-way ANOVA, grouped by host species: F = 0.174, P = 0.698 Two-way ANOVA, grouped by host sclerotization: F = 34.001, P = 0.004  74  T.A.M. NOTUM NOTUM-AS.1  AS 1 AS 1-2 AS 2 AS 2-3  AS 3 AS 3-4 AS 4 AS 4-6  C. bifida  non-flying  AS 6. Wing (vein)  13,33 14,21  Mites Bugs  Wing (memb.)  0  60^  100^20  40  60  80  100  AVG. MITES ATTACHED PER REGION (PERCENT OF TOTAL)  Figure 8. Experiment 2. Location of mite attachment on non-flying hosts at moderate salinity (10 000 to 15 000 )IS cm ') plotted as a percent of total mites collected on each host type. Numbers of mites in boxes are attached mites only (per replicate) associated with number of bugs directly below. Error bars are standard error. -  PARASITISM RATES (AVG. OF 2 REPS)  HOST  C. bifida^C. expleta  non-flying^non-flying  55.9^(+1 - 2.90) PREVALENCE^(+/-S.E.) PREVALENCE (ATTACHED ONLY)^(+/ - S.E.) 46.4^(+/-1.22) ABUNDANCE^(+/ - S.E.) ABUNDANCE (ATTACHED ONLY) (+/-s.E.) % OF MITES ATTACHED ON T.A.M.^(+/-s.E.)  Statistical Analysis between Hosts  t^P  69.3^(+1 - 0.94)  -1.230  0.435  62.6^(+/ - 1.73)  -1.760  0.329  2.10^(+/-0.24)  2.2^(+/-0.07)  -0.250  0.844  1.26^(+/-0.12)  1.4^(+/-0.14)  -2.077  0.286  50.3^(+/-1.48)  13.6^(+/-1.6)  145.4  0.004  Table 6. Experiment 2. Parasitism rates and statistical analyses for quantity and location of mites associated with non-flying hosts over 24 hours.  75 T.A.M. NO7UM (CT UM-AS . 1 As  -  ^I.  AS 14 AS 2 AS 2-3 AS 3 AS 3-4 AS 4 AS 4-5  flying^  flying  Wing (vein) Wing (memb.)  C. expleta  C. bifida  AS 5•  13,17 26,26  Mites Bugs  a..  1  40  0^20  13,18 33,29  Mites Bugs 1  60^80  100  20^  40  80^100  60  AVG MITES ATTACHED PER REGION (PERCENT OF TOTAL)  Figure 9. Experiment 3. Location of mite attachment on flying host types at moderate salinity (10 000 to 15 000 iS cm -1 ), plotted as a percent of total mites collected on each host type. Numbers of mites in boxes are attached mites only (per replicate) associated with number of bugs directly below. Error bars are standard error.  HOST  PARASITISM RATES AVG. OF 2 REPS) PREVALENCE (+/ S.E.) PREVALENCE (ATTACHED ONLY)^(+/  C. bifida^C. expleta flying^flying  21.2^(+/  -  -  S.E.)  ABUNDANCE^(+/-s.E.) ABUNDANCE (ATTACHED ONLY) (+/ S.E.) % OF MITES ATTACHED ON T.A.M.^(+/-s.E.) -  17.3^(+/  3.40)  32.8^(+/  3.40)  22.0^(+/  -  -  -  0.79^(+/  0.57^(+/ - 0.02)  0.51^(+/  78.5^(+/  52.3^(+/  4.85)  t^P  0.316  2.91)  0.909  0.530  0.02)  -1.214  0.439  0.04)  5.000  0.126  8.7)  5.798  0.109  -  -  Statistical Analysis between Hosts -1.849  -  0.69^(+/-0.00)  -  —  -  5.5)  Table 7. Experiment 3. Parasitism rates and statistical analyses for quantity and location of mites associated with flying hosts over 24 hours.  76 T.A.M. NOT UM NOTUM-AS1 AS. 1 AS 1 - 2 AS.2 AS. 2-3 AS 3 AS 3-4 AS 4 AS 4-6  C. expleta non-flying  AS. 6• Wing^(vein)  Mites Bugs^  Wing^(nemb.)  0  ^  20  ^  40  ^  60  C. expleta  flying  41,23  1  ^  19,11 28,22  Mites Bugs  26,22  80  ^  100^20^40^60^80^100  AVG. MITES ATTACHED PER REGION (PERCENT OF TOTAL)  Figure 10. Experiment 4. Location of mite attachment on C. expleta (non-flying and flying) at moderate salinity (10 000 to 15 000 pS cm -1 ), plotted as a percent of total mites collected on each host type. Numbers of mites in boxes are attached mites only (per replicate) associated with number of bugs shown directly below. Error bars are standard error.  PARASITISM RATES (AVG. OF 2 REPS) PREVALENCE (+/-s.E. PREVALENCE (ATTACHED ONLY)^(+/  48.9^(+/-4.43)  )  -  S.E.)  ABUNDANCE^(+/-s.E. ) ABUNDANCE (ATTACHED ONLY) (+/ S.E.) % OF MITES ATTACHED ON T.A.M.^(+/ 8 E.) -  -  HOST C. expleta^C. expleta non-flying^flying  44.8^(+1  -  4.61)  1.55^(+/ 1.31^(+/ 18.8^(+/  -  40.5^(+/-1.70) . 27.1^(+/-3.27)  Statistical Analysis between Hosts  t^P  0.326  0.799  1.049  0.485  -  0.08  1.48^(+/  -  0.05)  -0.109  0.931  -  0.09  0.59^(+/  -  0.03)  3.944  0.227  2.05)  75.6^(+1-1.1)  -57.08  0.011  Table 8. Experiment 4. Parasitism rates and statistical analyses for quantity and location of mites associated with C. expleta (non-flying and flying) over 24 hours.  77 B. Initial Mite Recruitment and Attachment: Low and High  Salinity 1. Low salinity. Question 3a (low salinity). ^Does low salinity affect quantity of initial mite recruitment? Data collected: High mortality of flying C. expleta in Experiment 5 did not permit analysis of this host type, but prevalence of mites at low salinity did not differ between the remaining three host types (F = 5.704, P = 0.095) (Table 9). Abundance followed the same trend. Comparing prevalences of these hosts at low salinity with the same hosts at moderate salinity (Experiment 1) revealed no effect attributable to salinity (Two-way ANOVA, grouped by salinity: F = 1.177, P = 0.320). Summary: Low salinity does not affect the quantity of initial mite recruitment (compared to moderate salinity).  Does low salinity affect quantity of initial mite attachment? Data collected: Prevalence (attached only) in Experiment 5 (Table 9) was not significantly different between host types at low salinity (F = 1.706, P = 0.320). Comparing the same host types at moderate salinity showed that there was no effect because of low salinity on prevalence (attached only) (Two-way ANOVA, grouped by salinity:^F = 1.320, P = 0.294).  78 Summary: Low salinity does not affect quantity of initial mite attachment (relative to moderate salinity).  Question 3b (low salinity). ^Does low salinity affect location of initial mite attachment? Data collected: Percentage of total mites attaching on the T.A.M. was greater on flying C. bifida compared to non-flying C. bifida and nonflying C. expleta (F = 110.65, P < 0.001) (Figure 11). 1 An analysis of the location of mite attachment on the three host types at moderate and low salinity showed no salinity effect (Two-way ANOVA, grouped by salinity: F = 3.701, P = 0.103). Summary: Low salinity did not affect the location of initial mite attachment (relative to moderate salinity).  2. High salinity: Question 3a (high salinity). Does high salinity affect quantity of initial mite recruitment? Data collected: Prevalence was significantly lower on non-flying C. bifida and non - flying C. expleta at high salinity (Experiment 6, Table  'The attachment of mites on one replicate of flying C. expleta is shown in Figure 11 for graphical comparison only. These data were not used for any statistical analysis.  79 10) compared to moderate salinity (Experiment 2, Table 6) (Twoway ANOVA, grouped by salinity: F = 94.794, P = 0.001). Athough mite recruitment was minimal at high salinity, it did occur. Abundance was also significantly lower at high salinity compared to moderate salinity. Summary: High salinity decreased the quantity of initial mite recruitment (with respect to moderate salinity). Does high salinity affect the quantity of initial mite attachment? Prevalence (attached only) was significantly less on hosts at high salinity (Table 10) than at low salinity (Table 6) (Two-way ANOVA, grouped by salinity: F = 87.199, P = 0.001). Abundance (attached only) followed the same trend. Summary: High salinity decreased the quantity of initial mite attachment (with respect to moderate salinity). Question 3b (high salinity). ^Does high salinity affect location of initial mite attachment? Very low attachment of mites at high salinity made this question unnanswerable.  80  ^ HOST PARASITISM ^ ^ ^ C. expleta^C. expleta RATES bifida C. bifida ^ C. ^ ^ flying non-flying^flying (AVG. OF 2 REPS) non-flying PREVALENCE (+/-s.E.) PREVALENCE (ATTACHED ONLY) (+/-s.E.) ABUNDANCE (+/-s.E. ) ABUNDANCE (ATTACHED ONLY) (+/-s.E.)  78.0  # of bugs  73.3 (+/ - 3.55)  37.1  67.8 (+/-^5.4)  37.1 (+/ - 0.38)  59.6 (+/-3.38)  alive  2.00  (+/ - 0.16)  1.50 (+/ - 0.18)  2.30  is too  1.45 (+/ - 0.12)  0.84 (+/ - 0.12)  1.76 (+/-0.21)  (+/ - 0.38)  (+1 - 2.33)  (+/ - 0.28)  low.  Table 9. Experiment 5. Parasitism rates for quantity of mites associated with all 4 host types over 24 hours at low salinity (< 300 pS cm -1 ).^All parasitism rates show no significant difference between hosts. ^For statistical analyses with moderate salinity (Experiment 1), see text.  81 Experiment 5. (Low salinity). T.A.M. NOTUM MOTUV-ASI  AS 1 AS 1-2 AS 2  I^ V^I  AS 2-3  AS. 3 AS 3-4  AS 4 AS. 4-6  C. bifida non-flying  AS 6• Wing (vein) Wing (memb.)  C. bifida flying  ^ Mites 22,21 ^ Bugs 12,19  Mites Bugs  30,10 25,21  T.A.M. NOTUM NOTUNFASI AS. 1  AS 1 - 2 AS. 2 AS. 2-3 AS 3 AS 3-4 AS 4 AS. 4 -6 AS. 6. Wing (vein) Wing (memb . )  C. expleta flying  C. expleta non-flying  I^  Mites  34,13 13,14  Mites Bugs  23 4  0^20^40^60^80^100^20^40^60^80^100 AVG MITES ATTACHED PER REGION (PERCENT OF TOTAL)  Figure 11. Experiment 5. Location of mite attachment on all host types at low salinity (< 300 pS cm -1 ), plotted as a percent of total mites collected on each host type. Numbers of mites in boxes are attached mites only (per replicate) associated with number of bugs shown directly below. Error bars are standard error. LOCATION OF ATTACHMENT as measured by percent of total mites on T.A.M. One-way ANOVA: F = 110.65, P = 0.001 Analysis of location of mite attachment between moderate salinity (Experiment 1) and low salinity (Experiment 5). Two-way ANOVA, grouped by salinity: F = 3.701, P = 0.103.  82  ^ ^ HOST Statistical Analysis PARASITISM ^ ^ ^ C. bifida C. expleta between Hosts RATES ^ ^ t^P (AVG. OF 2 REPS) non-flying^non-flying PREVALENCE^(%)(4/-s.E.) 13.9^(+/ PREVALENCE^(%) (ATTACHED ONLY) (+/ S.E.) 9.4^(+/ -  ABUNDANCE^(+/ S.EJ ABUNDANCE (ATTACHED ONLY) (+/  0.25)  10.9^(+/  3.50)  10.9^(+/  -  -  0.14^(+/  -  -  S.E.)  0.10^(+/  2.72)  -0.713  0.616  2.72)  -0.177  0.899  -  -  -  0.01  0.14^(+/  0.05)  -0.104  0.934  -  0.03  0.13^(+/-0.04)  -0.298  0.816  -  Table 10. Experiment 6. Parasitism rates and statistical analyses for quantity of mites associated with non-flying hosts over 24 hours at high salinity (> 18 000 pS cm '). For statistical comparison with moderate salinity (Experiment 2), see text. -  83 C. Long term Attachment and Engorgement. -  1. Attachment Data for this section were collected from Experiment 7 and were the summation of 8 days of data. The mortality data for the two host types, while not of direct importance to the questions asked below, are displayed in Table 11. This shows the dynamics of the mite-corixid interaction over 6 to 8 days on the predominant morphs of both host species. It was found in this study and in a similar study in 1990, that the mortality rate of non-flying C. expleta was higher in the first 3 days than flying C. bifida, but that this was true of unparasitized controls as well. Non-flying C. expleta was dying more quickly than flying C. bifida even without parasitism by E. euryhalina. As with questions 1, 2, and 3, recruitment followed the same trends as attachment. For brevity's sake, I do not mention recruitment in the Results of Questions 4 and 5, but consider it fully in the Discussion. Question 4a: Does host affect the quantity of mite attachment over 8 days? Data collected: In Experiment 7 (Table 12), prevalence (attached only) of E. euryhalina was significantly greater on non-flying (nonsclerotized) C. expleta than flying (sclerotized) C. bifida (t = -3.996, P = 0.028). Abundance (attached only) was also significantly different.  84 Summary: Over 6 to 8 days, non-flying C. expleta had significantly more mites attaching to it than flying C. bifida. Host type affected mite attachment. Question 4b: Does host affect location of mite attachment over 8 days? Data collected: The percentage of total mites attached on the T.A.M. in Experiment 7 was significantly greater on flying C. bifida than non-flying C. expleta (t = 6.319, P = 0.008) (See Figure 12). Mites on flying C. bifida were exclusively attached to the T.A.M. and wings, whereas mites on non-flying C. expleta were attached all over the dorsum, especially terga 2, 2-3, and 3. A X 2 analysis supported these data. Mites on flying C. bifida were attached more than would be randomly expected to the left and right thirds of the host dorsum, compared to the centre third (X  2  = 58.3, P < 0.001) Mites on non-flying C. expleta were attached more on the centre of the dorsum, compared to the left and right thirds (X 2 = 80.6, P < 0.001). Summary: Host type affected location of mite attachment over 8 days.  2. Mite Engorgement. Question 4a. Does host affect quantity of mite engorgement over 6 to 8 days? Data collected:  85 Prevalence (engorged only) in Experiment 7 was significantly greater on non-flying C. expleta compared to flying C. bifida in Experiment 7 (t = -3.83, P = 0.031) (see Table 13). Abundance (engorged only) was also significantly different. Summary: Host type affected the quantity of mites able to engorge over 6 to 8 days. Does host type affect location of mite engorgement over 6 to 8 days? Data collected: In Experiment 7, percentage of total mites collected that engorged on the T.A.M. was significantly greater on flying C. bifida compared to non-flying C. expleta (t = 64.2, P < 0.001) (Figure 13). All mites that commenced engorgement on flying C. bifida were found on the T.A.M., whereas most of the mites engorging on non-flying C. expleta were found on the centres of abdominal segments 2, 2-3, 3, and 3-4. A X 2 analysis of the mites on the left, right, and centre thirds of the host showed the same pattern as for mite attachment (X 2 c. b i fida = 37.2, P < 0.001, X 2 C. expleta 80.6, P < 0.005). In general, location of mite engorgement was very similar to mite attachment over 8 days. Summary: Host type affected the location of mite engorgement over 6 to 8 days.  ^1  Time (days) after initial infection ^ 0 - 8* 0 - 6 2.5^2.5 - 4.5^4.5 - 6 ^ 23.3 17.2 Percentage of C. bifida^3.96^7.76^5.50 ^ (+/- 3.92) dead (+/- S.E.)^(+/- 1.04)^(+/- 0.75)^(+/- 1.52) (+/- 1.55) ^ Percentage of total^4.49^26.1^8.92 59.8 39.6 ^ (+/- 4.87) mites collected (+/- S.E.) (+/- 2.50) ^(+/- 11.8)^(+/- 3.30) (+/- 8.13) Percentage of C. expleta^42.2^26.9^16.8 dead (+/- S.E.)^(+/- 2.74)^(+/- 4.12)^(+/- 3.84) Percentage of total^56.3^23.8^10.4 mites collected (+/- S.E.) (+/- 4.57) ^(+/- 3.37)^(+/- 3.66)  ^ 91.8 86.0 ^ (+/- 5.76) (+/- 4.52) ^ 89.6 90.5 ^ (+/- 7.33) (+/- 6.82)  Only 2 replicates were allowed to run for 8 days. Table 11. Experiment 7. Mortality data for flying C. bifida and non-flying C.expleta over 6 to 8 days of mite growth in Long Lake.  87 T.A.M.  NoTum NOVU110-AS1 As AS 1-2 AS 2 AS 2-3  =SW  AS 3  =MC  AS 3-4 AS 4 AS 4-6  Wing^(vein)  C. expleta  C. bifida  non-flying  flying  AS 6• ^  1  Wing^(memb.)  Mites Bugs  6,17,74,17  Mites Bugs  51,59,58,51  20^40^60^80^100  24,160,173,145 49,54,52,45  20^40^60^80  AVG. MITES ATTACHED PER REGION (PERCENT OF TOTAL)  Figure 12. Experiment 7. Location of mite attachment on flying C. bifida and non-flying C. expleta over 6 to 8 days at moderate salinity (10 000 to 15 000 pS cm '). Data are plotted as a percent of total mites collected on each host type. Numbers of mites in boxes are attached mites only (per replicate) associated with number of bugs shown directly below. Error bars are standard error. -  PARASITISM RATES AVG. OF 4 REPS  HOST C. bifida^C. expleta flying^non-flying  25.1^(+/ PREVALENCE (+/-S.E.) PREVALENCE (ATTACHED ONLY)^(+/- S.E.) 22.8^(4 ABUNDANCE^(+/-S.E.) ABUNDANCE (ATTACHED ONLY) (+/ - S.E.)  % OF MITES ATTACHED ON T.A.M.^(+/-S.E.)  -  Statistical Analysis between Hosts  2.65)  64.6^(+/-5.0)  -4.021  0.028  2.45)  63.7^(+/  0.21)  2.72^(4  -  5.0)  -3.996  0.028  -  0.31)  -3.241  0.048  0.51^(+/-0.11)  2.49^(+/-0.29)  -3.388  0.043  87.9(+/-3.15)  7.2^(+/-0.32)  6.319  0.008  0.71^(4  -  -  Table 12. Experiment 7. Parasitism rates and statistical analyses for quantity and location of attached mites associated with flying C. bifida and non-flying C. expleta over 6 to 8 days mite growth in Long Lake.  ^C.  88 T . A .M KA UM KIT UM- A.S.1 AS AS 1-2 AS.2 AS. 2 - 3 AS. 3 AS 3-4 AS. 4 AS. 4-6  C. bifida  flying  AS. Wing (vein)  MitesI 9 ,16, 38 ,13 ^I Bugs^51,59,58,51  Wing (memo.)  0  20  40  60  80  40^60^80^100  AVG. MITES ENGORGING PER REGION (PERCENT OF TOTAL)  Figure 13. Experiment 7. Location of mites engorging on flying C.  bifida and non-flying C. expleta over 6 to 8 days at moderate  salinity (10 000 to 15 000 pS cm-1 ). Data are plotted as a percent of total mites collected on each host type. Numbers of mites in boxes are attached mites only (per replicate) associated with number of bugs shown directly below. Data are plotted as a percent of total mites collected on each host type. Error bars are standard error.  ^  bifida  C.  HOST Statistical Analysis ^ expleta ^ between Hosts t^P flying^non-flying  PREVALENCE (ENGORGING ONLY)(+/-s.E.) 20.1^(+/ 2.36) 39.6 (+/- 2.66 ABUNDANCE (ENGORGING ONLY)(+/-s.E.) 0.33^(+/- o.o5) 0.86 (+/- 0.17) % OF MITES ATTACHED 13.3 (+/ 2.00) 100^(+/-0.0) ON T.A.M.^(+/ S.E.) -  -  -  -3.83  0.031  -3.47  0.040  64.2  <0.001  Table 13. Experiment 7. Parasitism rates and statistical analyses for quantity and location of engorging mites associated with flying C. bifida and non-flying C. expleta over 6 to 8 days mite growth in Long Lake.  89 Unfortunately, low sample sizes prohibited a statistical analysis of size of engorgement between host types over 6 and 8 days. Some qualitative data however, are given in Appendix 1 on the engorgement process of E. euryhalina.  D. Field Studies. Recall from the Materials and Methods that corixids were collected for 3 reasons: 1) To ensure that the quantity and location of E. euryhalina in laboratory experiments is representative of field infections (question 5). 2) To determine the relative proportions of C. bifida and C. expleta in lakes (i.e. that mites are still excluding C. expleta at low salinity). 3) To determine the relative proportions of flying and nonflying Cenocorixa spp. (i.e. that flying hosts predominate where mites are present).  1) Comparison of field parasitism and laboratory parasitism. Question 5a: Are laboratory experiments representative of the quantity of mites attaching on field collected corixids? Data collected: The September 13, 1991 collection had prevalence (attached only) on flying C. bifida of 29.4 % (Table 14b), on non-flying C. expleta of 46.8 % (Table 15b), and on C. expleta flying of 37.8 % (Table 16b), but these values were not significantly  90 different (F = 1.226, P = 0.297). In the collections of August 17, 1991 (Table 15a, 16a) and October 21, 1990 (Tables 14c, 15c, 16c), prevalence (attached only) was also not significantly different between host types.^Abundance (attached only) followed the same trend.^Collections of non-flying C. bifida were too low for statistical comparison. Summary: Prevalence of attached mites was not significantly different between host types in field collections. This result was the same as initial laboratory experiments, but different than the 8 day study between flying C. bifida and non-flying C. expleta which found a significant difference.  Question 5b: Are laboratory experiments representative of the  location of mites attaching on field collected corixids? Data collected: In the September 13, 1991 collection, there was no significant difference in the percentage of mites attaching on the T.A.M. between flying C. bifida (Figure 14), non-flying C. expleta (Figure 15) and flying C. expleta (Figure 16)^(F = 0.112, P = 0.894).^This result was different than lab experiments which showed significantly more mites on the T.A.M. of flying hosts compared to non-flying. Other field collections also found no significance in the location of mite attachment between hosts. Comparing individual host types collected in the field to  91 laboratory experiments, the attachment of mites on non-flying C. expleta was similar in that mites were mainly attached to abdominal segments 2, 2-3 and 3 (Figure 15). Attachment of mites to flying C. expleta was different in that a much lower proportion of mites was attaching to the T.A.M. in the field (Figure 16). Attachment in the field occurred through the hardened terga of abdominal segments 2, 2-3 and 3. Attachment of mites on flying C. bifida was similar to laboratory data in the July 17, 1991 collection with over 70 % attachment to the T.A.M., but different on September 13, 1991 and October 21, 1990 with less than 10 % of mites on the T.A.M. in each collection (Figure 14). Summary: Location of attachment was similar in the laboratory and the field on non-flying C. expleta, but different on flying C. expleta. On flying C. bifida, location of attachment was similar in one collection, but different in two others.  92  Field Samples T . A .M. MOTUM  July 17, 1991 Long Lake  6CITUM-A.S.1 AS 1 AS 1-2  AS 2  9  AS 2-3  ^ PARASITISM DATE RATES^July 17, 1991  AS. 3  9  AS. 9-4  PREVALENCE (%) PREVALENCE (t) (ATTACHED ONLY).  AS. 4  AS. 4-6  C. bifida  flying  AS 6•  Wing (vein)  ^ Mites ^26 Bugs 14  Wing (memb I  ABUNDANCE ABUNDANCE (ATTACHED ONLY)  Table 14a.  T.A.M.  57.1 57.1 3.07 1.86  ocaum September 13, 1991 Lye Lake  143TUA4-AS.1 AS 1 AS  1-2  AS.2 AS2-3  ^ PARASITISM DATE ^ RATES Sept. 13 1992  AS3 AS 3-4  PREVALENCE (%) PREVALENCE (t) (ATTACHED ONLY)  AS 4 AS 4-6  C. bifida  flying  AS 6.  Wing (vein) Wing (memb.)  ABUNDANCE ABUNDANCE (ATTACHED ONLY)  10 34  1 Mites Bugs  32.3 29.4 0.32 0.29  Table 14b. TAM NOTUM  October 21, 1990 Lye Lake  NOT UM-AS 1 AS.1 AS 1 -2 /2  ^ PARASITISM DATE ^ RATES Oct. 21,1990  AS2-3/3 •  AS.3-4/4  •  AS4-5/5+  PREVALENCE (%) PREVALENCE (%) (ATTACHED ONLY)  C. bifida  flying  WING  ^I ^  ABUNDANCE ABUNDANCE (ATTACHED ONLY)  ^ Mites Bugs 60 15  HEAD SPACE  0  20^40^60  80  100  25.0 25.0 0.25 0.25  Table 14c.  MITES ATTACHED PER REC*ON EFERCENTT OF TOTAL)  Figure 14. Field studies. Location of mite attachment on flying C. bifida by collection date, plotted as a percent of total mites collected. Table 14. Field studies. Quantity of mite attachment on flying C. bifida a.) July 17, 1991: Long Lake b.) September 13, 1991: Lake Lye c.) October 21, 1990: Lake Lye.  93  Field Samples  _  T.A.M. -^ MEnU1.4 -^  -  -^August 17,^1991^_  KrL1.4-As1  Lye Lake  -^  AS.1  -  -^  -  AS 2 -^  -  AS 1-2  AS 2-3 -^  I^  -^ I^  AS9  AS 3-4 =^ AS 4-6  _  C. expleta^-  -^  -^non-flying^_  AS 6.  Wing^(vein)  -  T.A.M.  I  Mites Bugs  Wing^(memb. 1 -  14 18  -  1^1^I  -2^  -  ktrum  PREVALENCE (%) PREVALENCE (%) (ATTACHED ONLY) ABUNDANCE ABUNDANCE (ATTACHED ONLY)  38.9 38.9 0.77 0.77  Table 15a.  -  Kraol- Aal AS 1  September 13, 1991^_ Lye Lake  AS1-2  -  §  PARASITISM DATE RATES^Aug. 17, 1991  -  AB2 ^ ^  LU^AS2-3  -  1-^AS. 3  -  2^  AS 34  -  AS4  -  AS4-6  -  C. expleta non-flying^-  AS 6• Wing^(vein) Wing^(mernb. I  I  Mites Bugs  30 47  1^i^1  -  -  PARASITISM DATE RATES^Sept. 13,1991  PREVALENCE (%) PREVALENCE (%) (ATTACHED ONLY) ABUNDANCE ABUNDANCE (ATTACHED ONLY)  48.9 46.8 0.66 0.64  Table 15b. TAM -  -  NCnliM -^October 21, 1990 Lye Lake WJM-ASI -  -  AS1 =  -  6 AS.1 - 2/2 =  -  IT AS2-3/3 -^  -  I  -  r- AS.3-4/4 -^I ASA-5/54 WING FEAR SPACE  -  -^  -  C. expleta  non-flying  I  Mites Bugs  14 ( 41^  -  i^I^t^I  0^20^40^60^80^100  mus ATTACHED PER REGION  PARASITISM RATES  PREVALENCE (%) PREVALENCE (%) (ATTACHED ONLY) ABUNDANCE ABUNDANCE (ATTACHED ONLY)  DATE Oct. 21, 1990 29.3 29.3 0.34 0.34  Table 15c.  PERCENT OF TOTAL)  Figure 15. Field studies. Location of mite attachment on nonflying C. expleta by collection date, plotted as a percent of total mites collected. Table 15. Field studies. Quantity of mite attachment on nonflying C. expleta a.) August 17, 1991: Lake Lye b.) September 13, 1991: Lake Lye c.) October 21, 1990: Lake Lye.  94 Field Samples T . A .M. OCTLA4  August 17, 1991  ►D71.114-AS.1  Lye Lake  AS 1 AS 1-2 AS. 2 AS 2-3  ^ DATE PARASITISM ^ Aug. 17, 1991 RATES  AS. 3 ti)  AS 9-4  PREVALENCE (%) PREVALENCE (%) (ATTACHED ONLY)  AS. • AS. 4-6  C. expl eta flying  As b•  1  Wing (vein) Wing (memb.)  T.A.M  i I  20^ 22^  Mites  Bugs  -  ABUNDANCE ABUNDANCE (ATTACHED ONLY)  45.5 31.8 1.09 0.91  Table 16a.  EJ  PCTLA4  September 13, 1991 Lye Lake  PCTLeA-AS 1 AS AS 1-2  6^  AS 2  W^AS 2-3  ^ PARASITISM DATE ^ RATES Se t. 13, 1991  AS. 3 AS 9-4 AS. 4  PREVALENCE (%) PREVALENCE (%) (ATTACHED ONLY)  0  AS 4-6  C. expleta flying  AS Wing (vein)  Mite s Bugs  Wing (memb.)  0  20^  40  24 45  60^80  100  ABUNDANCE ABUNDANCE (ATTACHED ONLY)  37.8 37.8 0.58 0.53  Table 16b.  MITES ATTACHED PEP REGION (PERCENT OF TOTAL)  Figure 16. Field studies. Location of mite attachment on flying C. expleta by collection date, plotted as a percent of total mites collected. Table 16. Field studies. Quantity of mite attachment on flying C. expleta a.) August 17, 1991: Lake Lye b.) September 13, 1991: Lake Lye.  95 2) Proportions of C. bifida compared to C. expleta. Percentages of C. bifida versus C. expleta in the various lakes over time are given in Table 17. C. bifida comprised over 90 % of Cenocorixa spp. in all collections in Long Lake (9 232 to 13 484 pS cm -1 ). Collections below this salinity in Near Opposite Crescent Pond, Greer Lake, and Box 27 never recorded C. expleta during 1990 or 1991, whereas C. bifida was present in all 3 bodies of water. In Lake Lye (9 807 to 13 493 pS cm -1 ), the percentage of C. expleta fluctuated from 57.2 % to 66.9 %. In the 2 high salinity lakes (Round-up and Barnes) (> 13 000 pS cm -1 ), C. expleta dominated comprising over 90 % of the total Cenocorixa spp.  96  Date of Collection Body of Water (salinity)  July 20/90  June 2/91  Long (Chilcotin) bif 94.6 % (9 232 to exp^5.4 % 13 484 pS cm -1 ) (n = 56)  bif 100 %  bif 100 %  (n = 32)  (n = 26)  Lye (9 807 to 13 493 pS cm')  n/a  bif 42.8 % exp 57.2 % (n = 28)  bif 33.8 % exp 66.2 % (n = 68)  bif 33.1 % exp 66.9 % (n = 148)  Barnes (13 687 to 19 528 pS cm 1 )  n/a  bif^4.2 % exp 95.8 % (n = 73)  n/a  n/a  Round-up (15 224 to 22 724 pS cm -1 )  n/a  bif^9.3 % exp 90.7 % (n = 54)  n/a  n/a  August 12/91  Sept.^14/91 bif 96.2 % exp^3.8 % (n = 26)  bif = C. bifida exp = C. expleta  Table 17. Field data. Percent composition of Cenocorixa spp. water boatmen collected in lakes by date. n/a = not available.  97 3) Proportions of flying and non-flying Cenocorixa spp. Percentages of non-flying (sclerotized and non sclerotized) -  and flying morphs are presented in Table 18. In Long Lake, the over-wintering population of C. bifida for 1990 was 100 % flying (based on June 2, 1991 sample: n = 32). The following overwintering generation (based on September 14th sample of 25) was 80 % flying. Lake Lye followed a similar trend with a June 2nd collection revealing 100 % flying (n = 12) and a September 14th collection of 77.3 % flying (n = 44). The few C. bifida that were found over-wintering at higher salinities (Barnes and Round-up) were also predominantly of the flying morph. C. expleta on the other hand was more often found as the nonflying morph. Lake Lye had 75.0 % non-flying (n = 16) on June 2nd and 69.6 % non-flying on September 13th (n = 92). The September collection had 18.5 % of the total C. expleta with entirely darkened dorsa, but only stage 2 wing muscle development (non-flying). (In June, 37.5 % of the C. expleta were non-flying, but completely dark.) Higher salinities had mostly non-flying C. expleta. Round-up had 77.6 % non-flying on June 2, 1991.  98  Date of Collection Body of Water^Corixid^June 2/91 August 12/91 Sept. 14/91 (salinity) ^Type Long (Chilcotin) bif nf^0 %^0 %^20.0 % 9 232 to^bif fly^100 %^100 %^80.0 % 13 484 pS crri l )^ (n = 32)^(n = 26)^(n = 25) exp nf (light)^n/a^n/a^0 0 exp nf (dark)^ 100 exp fly^ (n = Lye (9 807 to 13 493 pS cm -1 )  bif nf bif fly  0 % 100 % (n = 12)  exp nf (light) 37.5 % exp nf (dark) 37.5^% exp fly 25.0 % (n = 16) Barnes (13^687 to 19 528 pS cm - ')  bif nf bif fly  0 % 100 % (n = 3)  % % % 1)  13 % 87 % (n= 23)  22.7 % 77.3 % (n = 44)  51.1^% 24.5 % 24.4^% (n = 45)  51.1 % 18.5 % 30.4 % (n = 92)  n/a  n/a  exp nf (light) 62.9 % exp nf (dark) 17.1^% exp fly 20.0 % (n = 70) Round-up^bif nf^0 %^n/a^n/a (15 224 to^bif fly^100 % 22 724 pS^ (n = 5) exp nf (light) 38.8 % exp nf (dark) 38.8 % exp fly 22.4 % (n = 49) bif bif exp exp  nf = C. bifida non-flying fly = C. bifida flying nf = C. expleta non-flying fly = C. expleta flying  Table 18. Field data. Percentages of flying and non-flying Cenocorixa spp. in lakes at given times. n/a = not available.  99 DISCUSSION A. Initial Mite Recruitment and Attachment (Moderate Salinity). 1. Effect of Host Species.  Question la. Does host species affect quantity of initial mite recruitment? Host species does not affect quantity of initial mite recruitment. Recruitment, as measured by prevalence and abundance, was not significantly different between host species (Tables 5, 6, and 7). This contradicts Hypothesis la. It was expected that there would be a difference in recruitment based on the studies of Smith (1977). The lack of a significant difference means that the exclusion of C. expleta is not because more mites are initially finding C. expleta compared to C. bifida.  Does host species affect quantity of initial mite attachment?  Host species does not affect the quantity of initial mite attachment. Attachment, as measured by prevalence (attached only) and abundance (attached only) did not differ significantly between host species (Tables 5, 6, and 7). This also contradicts Hypothesis la. The exclusion of C. expleta is not because more mites are initially able to attach to C. expleta compared to C. bifida.  100 Question lb. Does host species affect location of initial mite attachment? Host species does not affect the location of initial mite attachment. Experiments with all hosts (Figure 7) and flying hosts (Figure 9) showed no significant difference in attachment between species when sclerotization was held constant. ^This supports Hypothesis lb.^In Experiment 2 (Figure 8) with nonflying hosts only, there appeared to be a species effect, as significantly more mites attached to the T.A.M. of non-flying C. bifida compared to non-flying C. expleta. Post-experimental  examination determined however, that some of these non-flying C. bifida had become partially sclerotized, resulting in inability  of the mites to attach to the abdominal segments. Since the location of mites does not differ between C.  bifida  and C.  expleta (when sclerotization remains constant), a species-  specific effect on mite location is not the cause of C. expleta's exclusion.  Overview of Initial Host Species Differences. I predicted Hypothesis la on mite recruitment and attachment because of the results of Smith (1977), who found significantly more mites attaching to C. expleta compared to C. bifida. Discrepancies between my results and his must have arisen because his experiments may have used both sclerotized and unsclerotized hosts, and were of variable time. Also, his low sample sizes and lack of replicates did not allow for the fact  101 that the attachment of E. euryhalina is largely dependent on chance, and variability within replicates and between host types is bound to occur. Acceptance of Hypothesis lb implies that there is no morphological or behavioural difference between C. bifida and C. expleta that would cause mites to attach in different locations  between species. Since the location of attachment is similar between species, energy loss because of parasitism should be the same. (Differences in energy loss are possible when mites become crowded on one host type: see Discussion, Question 2b.) The implication that morphological differences between species are not great enough to cause differences in mite parasitism is confirmed by Jansson (1972). His descriptions show that morphological differences between Cenocorixa bifida and C. expleta are primarily restricted to the hairs on the distal segment of the forelegs (Jansson, 1972). Since this area is not used for attachment by water mites, the lack of host differences in mite location and quantity is understandable. The only noticeable morphological difference that could cause differences in the initial mite attachment and/or recruitment is the predominance of the sclerotized morph of C. bifida compared to the normally unsclerotized C. expleta (see Question 2). Behavioural differences between species were similarly not great enough to cause differences in mite attachment. Davids (1973) speculated that differences in the grooming behaviour of corixids may account for differences in the quantity and  102 location of the mite Hydrachna conjecta. I saw both C. bifida and C. expleta groom against mites but could discern no difference between species. Another behaviour that might have caused higher quantity of mites on C. expleta, is if C. expleta spent more time near the water surface where mites collect. I observed that both species spent approximately the same amount of time near the water surface, and Reynolds (1974) concluded that habitat utilization in the field was not significantly different between species.  2. Effect of Host Sclerotization. Question 2a. Does host sclerotization affect quantity of initial mite recruitment? Host sclerotization does not affect quantity of initial mite recruitment. C. expleta's predominant unsclerotized morph did not have more mites recruiting to it than the predominant flying morph of C. bifida (Tables 5 and 8). This contradicts Hypothesis 2a. Greater recruitment to the predominant unsclerotized morph of C. expleta, compared to the predominant sclerotized morph of C. bifida, is not the cause of C. expleta's exclusion. Does host sclerotization affect quantity of initial mite attachment? Host sclerotization does not affect quantity of initial mite attachment. Initial attachment of E. euryhalina was not significantly different between sclerotized and unsclerotized  103 hosts (Tables 5 and 8). This also contradicted Hypothesis 2a. The exclusion of C. expleta is not because more mites are initially attaching to the predominant unsclerotized morph of C. expleta compared to the predominant sclerotized morph of C. bifida. Question 2b. Does host sclerotization affect location of initial mite attachment? Host sclerotization does affect location of initial mite attachment (Figures 7 and 10). Most mites on sclerotized morphs attached to the T.A.M., whereas mites on non-sclerotized morphs were attached mostly on abdominal segments 2 and 3. This supported Hypothesis 2b. The crowding of mites on the T.A.M. of sclerotized C. bifida should offer partial protection from the effects of mite parasitism compared to non-sclerotized C. expleta. This is part of the cause of the exclusion of C. expleta (see below).  Overview of Initial Host Sclerotization Differences. I predicted Hypothesis 2a on the recruitment and attachment of mites because I believed that differences in host sclerotization would cause differences in the quantity of mite attachment between host types, even after only 24 hours. reject Hypothesis 2a because I conclude that these differences are not evident in the early stages of parasitism (but see Question 4). Acceptance of Hypothesis 2b on the location of mites implies  104 that differences in the location of mite attachment between sclerotized and non-sclerotized hosts could be part of the cause of C. expleta's exclusion. In the first 24 hours of the mitecorixid interaction, it is the only factor that is important to the exclusion of C. expleta. The cause is that mites are more crowded on the T.A.M. of sclerotized hosts, compared to the abdominal segments of non-sclerotized hosts. There are many examples in which mites that are crowded together do not affect their hosts as much as non-crowded mites. Reilly and McCarthy (1991) found that Hydrachna conjecta were significantly smaller when 2 mites attached to the same hemielytron of their corixid hosts compared to when 1 mite attached to each hemielytra. (Smaller mite size is an indication of less host energy loss because of mite parasitism; Davids, 1973). Blockage of a mite's feeding tube or stylostome may be one cause of the lessening of energy drain from a host. This can occur through host defenses or through competition by the stylostomes of other mites (Abro, 1982). In interactions with relatively small hosts and proportionately large mites, impedance of mite growth can occur with even 2 mites per host. Aiken (1985) found significantly smaller Eylais sp. on the beetle Dytiscus alaskanus J. BalfourBrowne when 2 mites were present compared to single infections. Similar results were found by Lanciani (1971b) on beetles and corixids, and Davies (1959) on black flies. In cases of extreme crowding, mite mortality can occur, which  105  must cause less energy loss to the host than if evenly spaced mites are engorging. Mitchell (1968) found up to 50 % mortality of the mite Arrenurus mitoensis Imamura and Mitchell when mite density approached 30 mites per segment of the host damselfly Cercion hieroglyphicum. Observations of mites attached to the T.A.M. of Cenocorixa spp. indicate that in the later stages of engorgement, mite mortality occurs such that only one mite per T.A.M. can proceed past the initial stages of engorgement. Lanciani (1971b) suggested that the size of mites on aquatic insects is limited by the size of the subelytral space and demonstrated this through a study of the mite Hydrachna stipata Lundblad on a backswimmer of the genus Notonecta. When attached on the outside of the hemielytra, H. stipata was significantly larger than when attached on the underside of the hemielytra. Based on the size of a fully engorged E. euryhalina (less than 2 mm dorsal diameter = 6.28 mm 2 ), full engorgement of even one mite could not occur if attached on one of the T.A.M (a triangular area of 0.5 mm by 0.25 mm = 0.06 mm 2 )  .  I have witnessed the  effects of attempted engorgement on the T.A.M. from mites collected in the field on non-flying C. expleta. The mites become very elongated, with the anterior portion of their abdomens stretched, while engorgement proceeds only in the posterior of the abdomen, where the depth of the subelytral space is greater. Such deformity would most probably slow mite growth and certainly preclude metamorphosis to a nymphochrysalid.^Thus the maximum energy drain on flying C.  106 bifida would be less than the equivalent of 2 fully engorged  mites (1 per side), and energy may even be drained at a slower rate than from a normally engorging mite. Fully sclerotized hosts would, therefore, have an advantage over their nonsclerotized conspecifics if the abundance of mites in the population was over 2. The parasitism rates in the field often exceed this level as seen in my study and the more extensive field studies of Smith (1977). For example, my July 17th collection in Long Lake, taken at peak time for larval E. euryhalina, showed an abundance of 3.07 mites per host on flying C. bifida, but an abundance (engorged only) of only 0.5 mites per host. The sclerotization appeared to be providing the flying hosts with some protection from super-parasitism by limiting the number of mites that could commence engorgment to 2 or less. This number is within the range that flying C. bifida can withstand, as shown by the presence of 2 nymphochrysalid shells on some sclerotized hosts in the spring.  B. Initial Mite Recruitment and Attachment: Low and High Salinity 1. Low salinity. Question 3a (low salinity). ^Does low salinity affect quantity of initial mite recruitment? Low salinity does not affect quantity of initial mite recruitment compared to moderate salinity. Mite recruitment was  107 the same at low salinity (Table 9) and moderate salinity (Table 5) for three of the host types. (High mortality of flying C. expleta at low salinity precluded an analysis of this host type.) This supports Hypothesis 3a. Based on this conclusion, Experiments 1, 2, 3, and 4 should be indicative of mite recruitment throughout the salinity at which C. expleta is excluded.  Does low salinity affect quantity of initial mite attachment? Similarly, low salinity does not affect quantity of initial mite attachment compared to moderate salinity. Mite attachment was the same at low salinity (Table 9) compared to moderate salinity (Table 5) for three host types. This also supports Hypothesis 3a. Experiments 1 through 4 are indicative of mite attachment at the salinity at which C. expleta is excluded.  Question 3b (low salinity).^Does low salinity affect location of initial mite attachment? Low salinity does not affect location of initial mite attachment (Figure 11) relative to moderate salinity (Figure 7). Location of mite attachment on three host types was not significantly different between moderate and low salinity. This supports Hypothesis 3b. Experiments 1 through 4 are indicative of the location of mite attachment throughout the salinity at which C. expleta is excluded.  108 Overview on the Effects of Low Salinity on Mite Parasitism. Hypothesis 3a and 3b were predicted because I believed that the effects of mites would be the same from low salinity all the way through to the highest salinity from which C. expleta is excluded.^Evidence for this is provided by Scudder et al. (1972).^They found no difference between C. bifida and C. expleta in the long-term ability to osmoregulate in lake water  from 730 to 12 200 pS cm -1 , From this, they concluded that both were physiologically freshwater insects. In terms of the exclusion of C. expleta, the verification of Hypothesis 3 shows that low salinity causes no change in C. expleta (i.e. in behaviour) that would cause more mites to attach to it compared to C. bifida. The increased mortality of flying C. expleta may indicate however, that this host type is more susceptible to initial osmoregulatory shock than other host types.  2. High salinity: Question 3a (high salinity).^Does high salinity affect quantity of initial mite recruitment? High salinity does affect the quantity of intial mite recruitment compared to moderate salinity. Initial mite recruitment was significantly lower at high salinity (Table 10) than at moderate salinity (Table 6). My result contradicts Hypothesis 3a, which predicted no mites would recruit, but since there is a significant decrease in recruitment, I conclude that high salinity does affect recruitment of E. euryhalina.  109 Does high salinity affect quantity of initial mite attachment? See conclusions above for mite recruitment.  Question 3b (high salinity). ^Does high salinity affect location of initial mite attachment? Owing to very low attachment of mites at high salinity, a comparison of the location of mite attachment at moderate salinity was not possible.  Overview on the Effects of High Salinity on Mite Parasitism. Hypothesis 3a regarding high salinity was predicted because I believed that the factor that limited mites at high salinity was their inability to recruit and/or attach. With respect to the exclusion of C. expleta, the significant difference in attachment between moderate salinity and high salinity experiments shows that there is a salinity between 10 000 and 18 000 pS cm -1 that mites lose their ability to attach at the same rate. At this salinity, mites cease to have a substantial effect on C. expleta, and consequently, C. expleta ceases to be excluded.  110  C. Long-term Attachment and Engorgement.  1. Attachment Question 4a: Does host type affect the quantity of mite attachment over 6 to 8 days? Host type affected the quantity of mite attachment over 6 to 8 days. Attachment was significantly greater on the predominant non-sclerotized morph of C. expleta compared to the predominant sclerotized morph of C. bifida (Table 12). This supported Hypothesis 4a. The exclusion of C. expleta from low salinity is caused, at least partially, by the higher number of mites that are able to attach to the predominant morph of C. expleta over 6 to 8 days of mite exposure.  Question 4b: Does host type affect location of mite attachment over 6 to 8 days? Host type affected the location of mite attachment over 6 to 8 days. Mites attached more on the T.A.M. of sclerotized C. bifida than the T.A.M. of non-sclerotized C. expleta (Figure  12). This supported Hypothesis 2b. The decreased area of susceptibility of sclerotized C. bifida gives them a competitive advantage over non-sclerotized C. expleta, when in the presence of mites.  111  2. Mite Engorgement. Question 4a. Does host type affect quantity of mite engorgement over 6 to 8 days? Host type affected the quantity of mite engorgement over 6 to 8 days. Significantly more mites were able to engorge on nonsclerotized C. expleta than on sclerotized C. bifida (Table 13). This supported Hypothesis 4a. After 6 to 8 days, the greater number of engorging mites on the predominant non-sclerotized morph of C. expleta is at least partially the cause of C. expleta's exclusion.  Question 4b. Does host type affect location of mite engorgement over 6 to 8 days? Host type affected the location of mite engorgement over 6 to 8 days. Significantly more mites engorged on the T.A.M. of flying C. bifida than on the T.A.M. of non-flying C. expleta (Figure 13). This supported Hypothesis 4b. After 6 to 8 days, the crowded nature of the engorging mites on flying C. bifida impedes some of the mites' growth. This effect gives partial protection to flying C. bifida, whereas non-flying C. expleta is fully affected by mites, and is excluded from areas where mites are prevalent.  112 Overview of Long-term Mite Attachment and Engorgement. Acceptance of Hypotheses 4a on mite attachment and engorgement is the first evidence I found that more mites are associated with C. expleta than C. bifida. Since I used the predominant field morphs in my long-term experiment, the attachment differences witnessed should be most representative of what is occurring to the two host species in the field. Higher prevalence on one host type has been shown to correlate with lower host fecundity (Davids and Schoots, 1975; Smith, 1977) and higher host mortality (Lanciani 1975, 1986). This could be the most important factor in the exclusion of C. expleta from lakes of low and moderate salinity.  The question of whether there is a long-term species effect or an effect of sclerotization was not fully determined because of the use of only the predominant morphs of each species. As already stated (pp. 60-61), I would have liked to have used all 4 hosts types as I did for Questions 1, 2, and 3, but use of teneral non-flying C. bifida over 6 to 8 days would have presented problems in analysis because some of them harden in this time. Nevertheless, the use of the predominant morphs of each species assured that the results in my experiments are representative of what is occurring in the field. From my finding that species has no effect initially, whereas sclerotization affects the location of initial mite attachment, I would speculate that the effects of sclerotization are more responsible for the results of Experiment 7.  113 I found that over 6 to 8 days, recruitment of mites (as measured by prevalence and abundance) differed between hosts. Since 24 hour studies (Experiments 1 through 4) indicated that mites recruit to hosts equally, differences in recruitment over 6 to 8 days must be because mites are leaving the hosts after 24 hours, or they are being removed by the hosts. I observed that disturbed E. euryhalina are capable of disattaching within three days of initial infection. These mites, perhaps failing to locate the T.A.M. or finding this location unacceptable, might leave the host and could conceivably even find and attach to a non-flying C. expleta. Mites were observed under laboratory conditions entering and exiting the air space of hosts after a few minutes of exposure. With respect to mites being removed, Harris and Harrison (1974) stated that Hydrachna sp. (H. barri by Smith, 1987) may be knocked off the legs of their corixid hosts, and Davids (1973) speculated that H. conjecta may be brushed of the underside of the hemielytra. Mites could be dislodged during flight and since the lake temperature during Experiment 7 was above the temperature at which flight is initiated (15 ° C according to Scudder, 1969a), flight in the covered enclosures was possible. Whatever the cause, over 6 to 8 days, more mites were found on non-flying C. expleta compared to flying C. bifida, despite similar initial recruitment. Mite attachment, as measured by prevalence (attached only) was also significantly lower on flying C. bifida (Table 12). Mite engorgement, as  114 measured by prevalence (engorging only), followed the same trend (Table 13). The process by which C. bifida lowers the number of potential attaching and engorging mites is an important factor in its ability to withstand mite parasitism. Non-flying C. expleta, failing to lower mite engorgement is excluded from  areas with mites. Acceptance of Hypothesis 4b, regarding location of mites over 6 to 8 days, causes me to reach the same conclusions as for the effects of sclerotization on initial mite attachment (Hypothesis 2b). Since I was using only two host types, I cannot, with certainty, attribute the difference in location of mite attachment to a host sclerotization effect rather than a species effect. It does, however, seem most plausible that sclerotization is causing the effect, because of the initial differences caused by sclerotization. The fact that 100 % of the mites engorging on flying C. bifida are on the T.A.M. shows that this is truly the only susceptible area on this host type. Based on previous arguments (see Discussion, Hypothesis 2), flying C. bifida is protected from mite parasitism significantly more than non-flying C. expleta.  115 D. Field Studies. 1) Comparison of field parasitism and laboratory parasitism.  Question 5a: Are laboratory experiments representative of the quantity of mites attaching on field collected corixids? Initial laboratory experiments were representative of the quantity of mites attaching on field collected corixids, but laboratory experiments over 6 to 8 days differed from field collections. In both initial laboratory experiments and field collections, I found no significant difference between host types. In 6 to 8 day laboratory experiments I did find a significant difference between non-flying C. expleta and flying C. bifida. Since it was found that attachment of mites in the field did not differ between host types, there is no evidence from these field collections that mites are affecting C. expleta more than C. bifida (but see Overview).  Question 5b: Are laboratory experiments representative of the location of mites attaching on field collected corixids? Laboratory experiments were representative of location of mite attachment on non-flying C. expleta, but not representative on either flying host. This supported Hypothesis 5b. Fieldcollected non-flying hosts are always non-sclerotized, providing mites with their preferred attachment sites on abdominal segments 2 and 3. Field-collected flying hosts however, may have had mite attachment prior, during, or after sclerotization,  116 creating high variability in the location of mite attachment. Considering this in relation to how sclerotization protects flying hosts (Question 2b), mites will have a much longer period  of time to find, attach and engorge on non-flying C. expleta compared to flying C. bifida. Location of mite attachment in the field supports the theory that water mites have a greater effect on C. expleta than on C. bifida.  Overview of Field Collections From field collections, there was not a significantly greater quantity of mites attaching to C. expleta compared to C. bifida. My 6 to 8 day laboratory experiment would indicate, however, that more mites do attach to non-flying C. expleta compared to flying C. bifida over 6 to 8 days. This contradiction requires an explanation. The field collections of Smith (1977), Reilly and McCarthy (1991), and Aiken (1985) show great variability in the quantity of Eylais spp. mites attaching to their hosts over time. They all show a relatively low overwintering prevalence, a higher peak in early summer when the summer generation of larval mites is present, and a decrease in prevalence shortly after; accounted for by completion of larval mite development, and/or death of infected hosts. My collections showed no difference in prevalence, but this does not preclude the fact that E. euryhalina is excluding C. expleta. If mites are recruiting to their hosts at the same  117 rate (Experiment 1 through 4), then no difference would be seen in field collections at this time. Also, if C. expleta are dying more quickly because of the later stages of engorgement, then a collection at this time might actually show lower parasitism on C. expleta, because only uninfected hosts are left in the population. My collections must have been at times in the mite-corixid interaction when prevalences did not differ between hosts. Both Smith (1977) and I found no nymphochrysalids on C.  expleta. This would suggest that C. expleta has a higher longterm mite-induced mortality which is supported by my findings of Experiment 7. This finding may be because of an effect of host species and/or host sclerotization. Acceptance of Hypothesis 2b on the location of mites in the field supports the theory that mites are excluding C. expleta because of the predominance of its non-flying morph. From my findings in the laboratory and the field, the sequence of events that creates a differential effect of parasitism on C. expleta is as follows: Larval E. euryhalina recruit to all hosts at the same rate (Table 5). If the potential host is non-sclerotized (nonflying), mites attach to the preferred sites in the middle of the abdominal dorsum (Figures 7, 13: non-flying hosts). Presumably, as sclerotization continues, there becomes a point at which the integument of the abdomen is too hard or thick for the mite's mouthparts to pierce, at which time attachment to the  118 T.A.M. becomes the mites' only option apart from leaving the host (Figures 7, 10, and 13: flying hosts). This exodus of mites from sclerotized hosts is reflected in significantly less attached and engorging mites on flying C. bifida compared to non-flying C. expleta over 6 to 8 days (Tables 12 and 13). As postulated in the discussion of question 2b, attachment to the T.A.M. probably does not cause as much energy loss to the host. Therefore, in terms of parasitism, it is in the host's best interest to become sclerotized as quickly as possible. Support for the hypothesis that host sclerotization controls mite location is derived from a comparison of studies of mites that commence engorgement on teneral hosts during eclosion, with studies of mites that begin engorging on sexually mature (sclerotized) insects. Mites on teneral hosts are found engorging directly through the sclerites. Examples are Hydrachna virella Lanciani on the dorsal pronotum of the backswimmer Buenoa scimitra Bare (Lanciani, 1980); Arrenurus agrionicolus Uchida on the ventrum of the 7th abdominal segment of the damselfly Cercion hieroglyphicum (Mitchell, 1968); and Hydryphantes tenuabilis Marshall on the marsh treader Hydrometra australis (= myrae) Drake and Hottes that attaches through the abdominal segments directly after host eclosion (Lanciani, 1971a). In contrast, the studies of mites that attach to hardened hosts show these hosts engorging exclusively through membranous regions. Limnochares americana Lundblad attaches to various  119 damselflies, but always at the bases of the legs (Conroy and Kuhn, 1977; life cycle from Smith and Cook, 1991). Partuniella thermalis usually attaches to the alary membranes at the base of the wings of the brine fly Paracoenia sp. (Wiegert and Mitchell, 1973). Thyas barbigera Viets attaches to the posterior face of the thorax of Aedes spp. mosquitoes 72.3 % of the time, and 9.3 % of the time on the cervical membrane between the head and thorax (Mullen, 1977). 2 Arrenurus spp. attach at the base of the legs of the damselfly Lestes sponsa (Hansemann) (Abro, 1982).  2. Proportions of C. bifida compared to C. expleta. Based on the field data from the summers of 1990 and 1991 (Table 17), it would appear that mite parasitism is still affecting C. expleta in the same way it did in the study of Smith (1977). In the summers of 1990 and 1991, E. euryhalina was breeding in all lakes up to and including the salinity of Lake Lye (9 807 to 13 493 pS cm - '). Below this salinity, C. bifida comprised over 90 % of the Cenocorixa fauna in all lakes, while above this salinity, C. expleta comprised over 90 % of the Cenocorixa fauna (Table 17). In Lake Lye, where mite prevalance was moderate (not as high as in Long Lake), C. expleta comprised 57.2 % to 66.7 % of the Cenocorixa spp. This lake is at the upper limit of E. euryhalina's salinity tolerance, and from my findings on the effects of high salinity on E. euryhalina (Table  2 The posterior face of the thorax of mosquitoes is one of the softer parts of the body (Corbet, 1963).  120 6 compared to Table 10), the salinity may be lessening the effects of mite parasitism.^This would allow C. expleta to establish itself. The field data of 1990 and 1991 on the occurrence of C. expleta, C. bifida and E. euryhalina support the hypothesis of Smith (1977) that parasitic water mites exclude C. expleta from lakes of low salinity.  3. Proportions of flying and non-flying Cenocorixa spp. The proportions of flying and non-flying Cenocorixa spp. during the summers of 1990 and 1991 were similar to those reported in previous studies (Table 3). C. bifida was predominantly flying, whereas C. expleta was predominantly nonflying (Table 18). Smith (1977) reports that in the overwintering populations of 1976 and 1977, the non-flying morph accounted for no more than 1 % of the population. In summary, the overwintering populations of C. bifida were predominantly flying in 1962, 1963, 1967, 1976, 1977, 1990, and 1991, or 7 out of 10 years studied. Only in 1968 and 1969 did the non-flying morph predominate, while in 1966, the flying and non-flying morphs were in equal proportions. In contrast, C. expleta was predominantly non-flying in the two years of my investigation, and all previous studies have found that C. expleta is mostly non-flying (Scudder, 1975). As previously stated, the difference in the percentage of flying morphs between C. bifida and C. expleta must affect the survivorship of the two species when field abundance levels are  121 2 mites per host or higher (Question 2b). Although both species begin their adult lives unsclerotized, C. bifida usually becomes flying and sclerotized, making it vulnerable only directly after eclosion to adult, a period which takes no more than 8 days at 20 ° C (Scudder, 1971). C. expleta, however, will usually remain non-flying for its entire adult life, meaning that it must avoid the gauntlet of parasites for its entire adult life. This difference in the temporal "window of opportunity" for water mites, is the basis of the exclusion of C. expleta at low salinity.  E. General Discussion My study of E. euryhalina has shown a greater effect of parasitism on C. expleta compared to C. bifida. The question remains, however, does a study of only E. euryhalina adequately account for the effects of all mites on Cenocorixa spp.? Other mites are present at Becher's Prairie, especially E. discreta and H. davidsi Smith, which are sometimes more prevalent than E. euryhalina. especially at low salinity. With respect to E. discreta, there appears to be no reason why sclerotization of the host dorsum would not affect attachment in the same manner as for E. euryhalina. Davids et al. (1977) described the preferred attachment sites in the field for E. discreta on some corixids of the. genera Sigara and Cymatia. They found that larval E. discreta preferred abdominal segments 2, 3 and 4 on all hosts and my observations on E. discreta  122 concur with these findings. Lanciani (1969) claims that E. discreta prefers the tergum of abdominal segment 3, followed by 2 and 4, with attachment to abdominal terga 1 and 5 only on heavily parasitized hosts. It appears that at low salinity, E. discreta and E. euryhalina occupy a similar niche, sometimes compete for the same attachment sites, and must be affected in the same way by host sclerotization. I expect that a fully sclerotized host would prevent attachment of E. discreta in its preferred site because of field observations of E. discreta attaching to the T.A.M. of flying hosts as well as dead, dessicated E. discreta on the hardened terga of flying hosts. Preliminary measurements of the thickness of a sclerotized abdominal tergum show that it remains very thin, even when sclerotized, indicating that it is not the thickness of the sclerite that protects the host from mite engorgement. Longer chelicerae, therefore, would not help E. discreta to engorge because the host defense is more likely based on the hardness of the integument rather than its thickness. Hydrachna davidsi poses another challenge to the exclusion theory based on sclerotization. ^Their attachment is to the underside of the hemelytra. As with Eylais spp. larvae, the parasitism only occurs on adult (winged) hosts, so in this respect, the effects of parasitism are acting on the same members of the population. I am not sure, however, that the degree of sclerotization of the hemelytra differs between flying  123 and non-flying hosts. If this were true, attachment of H. davidsi should be possible on both C. bifida and C. expleta for  the entirety of their adult lives, irrespective of the degree of sclerotization. From this assumption, H. davidsi could not be limiting C. expleta by the same mechanisms as Eylais spp., if at all. Attachment of H. davidsi on C. expleta and C. bifida was studied by Smith (1977). He found that mites attached significantly more often (P < 0.05) on C. expleta than on C. bifida. The mechanism responsible is not known, and a full  study on H. davidsi and the sclerotization of the hemelytra of Cenocorixa spp. would be required to determine this. Nevertheless, the absence of C. expleta in low salinity lakes can be explained without an understanding of the effects of H. davidsi. The high prevalence of E. discreta and presence of E.  euryhalina at low salinity means that there is no water body below the salinity of Barnes Lake (13 687 to 19 528 pS cm -1 from my study), in which parasitism will not occur on C. expleta. Combining the mite parasitism data from the lab and field with the historical percentages of flying and non-flying Cenocorixa spp., one can make speculate on the co-evolution of mites and water boatmen. The long-term fluctuations in flying morphs of C. bifida are explained by differences in the temperature of the lake during development of the last larval instar (Scudder and Meredith, 1972). Above 15 ° C, development to the flying morph occurs in  124 both species in the laboratory and interspecific differences in the percentage of flying morphs are thought to arise from slight differences in the timing of their life cycles (Jansson and Scudder, 1974). It is possible, however, that abiotic factors are not the only forces that control the proportions of flying morphs found in the field. The hypothesis is forwarded here, that parasitism by water mites is one factor accounting for differences in the composition of corixid populations. The findings of this work show a greater effect of parasitism on non-flying individuals compared to flying forms. Increased mortality of non-flying individuals in regions of parasitism (low to moderate salinities) would cause the percentage of nonflying individuals to be lower than in areas with no parasites (high salinity). This helps explain the finding of Scudder (1975) that at the lower salinity range of each species, there are greater percentages of flying morphs than at the higher limits of the salinity range. In addition, parasitism may act as a selection agent favouring flying morphs. One would expect that in areas of high parasitism, both species would favour the production of the flying morph at lower and lower temperatures towards a physiological minimum. This minimum may be 15 ° C as reported by Scudder and Meredith (1972). In contrast, in permanent lakes where there are no parasites, there may be selection pressure against production of the flying morph (with its associated ovarian diapause), because it is associated with reduced  125 reproductive fitness (Young, 1965b). There would be no reason for an individual in a permanent, productive lake to forego reproduction so that it could disperse. In these lakes, the percentage of flying morphs should be lower, as reported in Scudder (1975), and the temperature of inducement of wing muscle development should be higher. From this reasoning, it can be further hypothesized that mite parasitism could have played a role in the speciation process of C. expleta and C. bifida. The ancestor of C. bifida and C. expleta would have had two different selection forces on it when  mite parasitism evolved. At low salinities, where mites were present, the population would have evolved into a sclerotized, migratory form. At high salinity, where mites were not present, the population would have evolved into a non-flying, nonsclerotized, but more fecund form. Then through reproductive isolation, these two subpopulations may have speciated into the present-day C. bifida and C. expleta.  126 REFERENCES Abro, A. 1982. The effects of parasitic water mite larvae (Arrenurus spp.) on Zygopteran imagoes (Odonata) J. Invert. Pathol. 39: 373-381. Aiken, R. B. 1985. Attachment sites, phenology and growth of the larvae of Eylais sp. (Acari) on Dytiscus alaskanus J. Balfour-Browne (Coleoptera: Dytiscidae). Can. J. Zool. 63: 267-271. Cannings, R. A. and S. G. Cannings. 1987. The Odonata of some saline lakes in British Columbia, Canada: ecological distribution and zoogeography. Adv. Odonatol. 3: 7-21. Cannings, R. A. and G. G. E. Scudder. 1978. The littoral Chimomidae (Diptera) of saline lakes in central British Columbia. Can. J. Zool. 56: 1144-1155. Cannings, S. G. 1978. Successful rearing of Cenocorixa expleta (Uhler) (Hemiptera: Corixidae) in fresh water. Can. Ent. 110: 559. Conroy, J. C. and J. L. Kuhn. 1977. New annotated records of Odonata from the province of Manitoba with notes on their parasitism by larvae of water mites. Manitoba Ent. 11: 2740. Corbet, P. S. 1963. Reliability of parasitic water mites (Hydracarina) as indicators of physiological age in mosquitoes (Diptera: Culicidae). Entomol. Exp. Appl. 6: 215-233. Crofton, H. D. 1971. A quantitative approach to parasitism. Parasitology 62: 179-193. Davids, C. 1973. The water mite Hydrachna conjecta Koenike, 1895 (Acari, Hydrachnellae), bionomics and relation to species of Corixidae (Hemiptera). Neth. J. Zool. 23: 363-429. Davids, C. and C. J. Schoots. 1975. The influence of the water mite species Hydrachna conjecta and Hydrachna cruenta (Acari, Hydrachnellae) on the egg production of the Corixidae Sigara striata and Cymatia coleoptrata (Hemiptera) Verh. Int. Ver. Limnol. 19: 3079 82. -  Davids, C., G. J. Nielsen and P. Gehring. 1977. Site selection and growth of the larvae of Eylais discreta Koenike, 1897 (Acari, Hydrachnellae). Bijdr. Dierkd. 46: 180-184.  127 Davies, D. M. 1959. The parasitism of black flies (Diptera, Simuliidae) by larval water mites of the genus Sperchon. Can. J. Zool. 52: 1155-1165. Fernando, C. H. and D. Galbraith. 1970. A heavy infestation of gerrids (Hemiptera--Heteroptera) by water mites (Acarina-Limnocharidae). Can J. Zool. 37: 353-369. Forbes, M. R. L. 1991. Ectoparasites and mating success of male Enallagma ebrium damselflies (Odonata: Coenagrionidae). Oikos 60: 336-342. Gledhill, T., J. Cowley and R. J. M. Gunn. 1982. Some aspects of the host:parasite relationships between adult blackflies (Diptera; Simuliidae) and larvae of the water-mite Sperchon setiger (Acari; Hydrachnellae) in a small chalk stream in southern England. Freshwat. Biol. 12: 345-57. Harris, D. A. and A. D. Harrison. 1974. Life cycles and larval behaviour of two species of Hydrachna (Acari, Hydrachnellae) parasitic upon Corixidae (Hemiptera-Heteroptera). Can. J. Zool. 52: 1155-1165. Jansson, A. 1972. Systematic notes and new synonomy in the genus Cenocorixa (Hemiptera:Ceorixidae). Can. Ent. 104: 449-459. Jansson, A. 1973. Stridulation and its significance in the genus Cenocorixa (Hemiptera, Corixidae). Behaviour 46: 1-36. Jansson, A. and G. G. E. Scudder. 1974. The life cycle and sexual development of Cenocorixa species (Hemiptera, Corixidae) in the Pacific Northwest of North America. Freshwat. Biol. 4: 73-92. Lancaster, J. 1985. Structure of arthropod communities in some saline lakes of central British Columbia. M. Sc. Thesis, Department of Zoology, University of British Columbia. 192 pp. Lanciani, C. 1969. Three species of Eylais (Acari:Eylaidae) parasitic on aquatic Hemiptera. Trans. Amer. Microsc. Soc. 89: 169-188. Lanciani, C. 1970a. Resource partitioning in species of the water mite genus Eylais. Ecology 51: 338-342. Lanciani, C. 1970b. New species of Eylais (Acari: Eylaidae) parasitic on aquatic Coleoptera. Trans. Amer. Microsc. Soc. 89: 169-188.  128 Lanciani, C. 1971a. Host exploitation and synchronous development in a water mite parasite of the marsh treader Hydrometra myrae (Heteroptera: Hydrometridae). Ann. ent Soc. Am. 64: 1254-1259. Lanciani, C. 1971b. Host-related size of parasitic water mites of the genus Eylais. Am. Midl. Nat. 85: 242-247. Lanciani, C. 1975. Parasite-induced alterations in host reproduction and survival. Ecology 56: 689 695. -  Lanciani, C. 1980. Parasitism of the backswimmer Buenoa scimitra (Hemiptera: Notonectidae) by the water mite Hydrachna virella (Acari: Hydrachnellae). Freshwat. Biol. 10: 527-532. Lanciani, C. 1983. Overview of the effects of water mite parasitism on aquatic insects. In Hoy, M. A., G. L. Cunningham, and L. Knutson (Eds.). Biological Control of Pests by Mites. Spec. Publs. Univ. Calif. agric. Exp. Stn. 3304. pp. 86-90. Lanciani, C. 1986. Reduced survivorship in Dasyhelea mutabilis (Diptera: Ceratopogonidae) parasitized by the water mite Tyrrellia circularis (Acariformes: Limnesiidae). J. Parasit. 72: 613-614. Lanciani, C. and J. M. Boyett 1980. Demonstrating parasitic water mite-induced mortality in natural host populations. Parasitology 81: 465-475. Lanciani, C. and P. G. May 1982. Parasite-mediated reduction in growth of nymphal backswimmers. Parasitology 85:1-7 McCrae, A. W. R. 1976. The association between larval parasitic water mites (Hydracarina) and Anopheles implexus (Theobald) (Diptera: Culicidae). Bull. Entomol. Res. 66: 633-650. Margolis, L., G. W. Esch, J. C. Holmes, A. M. Kuris and G. A. Schad. 1982. The use of ecological terms in parasitology (report of an ad hoc committee of the American Society of Parasitologists). J. Parasit. 68: 131-133. Martin, N. A. 1975. Observations on the rationship Eylais and Hydrachna (Acari: Hydracarina) and Sigara spp. (Insecta: Hemiptera: Corixidae). N. Z. J. Zool. 2: 45-50. Minchella, D. J. and M. E. Scott 1991. Parasitism: A cryptic determinant of animal structure. Trends Ecol. Evol. 6(8): 250-254.  129 Mitchell, R. D. 1967. Host exploitation of two closely related water mites. Evolution, Lancaster, Pa. 21: 59-75. Mitchell, R. D. 1968. Site selection by larval water mites parasitic on the damselfly Cercion hieroglyphicum Brauer. Ecology 49: 40-47. Mullen, G. R. 1977. Acarine parasites of mosquitoes. IV. Taxonomy, life history and behaviour of Thyas barbigera and Thyasides sphagnorum (Hydrachnellae: Thyasidae). J. Med. Ent. 13. 475-485. Nielsen, G. J. and C. Davids. 1975. Contributions to the knowledge of the morphology of the larvae of four European Eylais species (Acari, Hydrachnellae). Acarologia 17: 519-528. O'Connor, B. M. 1984. Phylogenetic relationships among higher taxa in the Acariformes, with particular reference to the Astigmata. In Griffiths, D. A. and Bowman, C. E. (Eds). Acarology VI (vol. 1). pp 19-27. Parsons, M. C. 1970. Respiratory significance of the thoracic and abdominal morphology of three aquatic bugs, Ambrysus Notonecta, and Hesporocorixa (Insecta, Heteroptera). Z. Morph. Tiere 66: 242-298. Piatakov, M. L. 1916. 0 nekotorykh Eylais i Hydrachna, parazitiruyushchikh pod nadkrylyami Dytiscidae i Gyrinidae (Acarina, Limnocharidae). Russk. ent. Obozr. 15: 125-130. Prasad, V. and D. R. Cook 1972. The taxonomy of water mite larvae. Mem. Am. ent. Inst. 18. 326 pp. Reilly, P. and T. K. McCarthy. 1991. Watermite parasitism of Corixidae: infection parameters, larval mite growth, competitive interaction and host response. Oikos 60: 137148. Reynolds, J. D. 1974. Aspects of the ecology of two species of Cenocorixa (Corixidae:Hemiptera) in allopatry and sympatry. Ph.D. thesis, University of British Columbia, Vancouver, Canada. Reynolds, J. D. 1979. Crustacean zooplankton of some saline lakes of central British Columbia saline lakes. Syesis 12: 169-173. Reynolds, J. D. and S. C. Reynolds. 1975. Aquatic angiosperms of some British Columbia saline lakes. Syesis 8: 291-295.  130 Scudder, G. G. E. 1966. The immature stages of Cenocorixa bifida (Hung.) and C. expleta (Uhler) (Hemiptera: Corixidae). J. ent. Soc. Br. Columbia 63: 33 40. -  Scudder, G. G. E. 1969a. The distribution of two species of Cenocorixa in inland saline lakes of British Columbia. J. ent. Soc. Br. Columbia 66: 32 41. -  Scudder, G. G. E. 1969b. The fauna of saline lakes on the Fraser Plateau in British Columbia. Verh. int. Ver. Limnol. 17: 430-439. Scudder, G. G. E. 1971. The postembryonic development of the indirect flight muscles in Cenocorixa bifida (Hung.) (Hemiptera:Corixidae). Can. J. Zool. 49: 1387-1398. Scudder, G. G. E. 1975. Field studies of the flight muscle polymorphism in Cenocorixa (Hemiptera:Corixidae). Verh. Int. Verein. Theor. Angew. Limnol. 19: 3064 3072. -  Scudder, G. G. E. 1976. Water-boatmen of saline water (Hemiptera: Corixidae). In: L. Cheng. (ed.), Marine Insects. North Holland Publ. Co., Amst. 263-289. Scudder, G. G. E. 1983. A review of factors governing the distribution of two closely related corixids in the saline lakes of British Columbia. Hydrobiologia 105: 143-154. Scudder, G. G. E. and K. H. Mann. 1968. The leeches of some lakes in the southern interior plateau of British Columbia. Hydrobiologia 105: 143-154. Scudder, G. G. E. and J. Meredith. 1972. Temperature-induced development in the indirect flight muscle of adult Cenocorixa (Hemiptera: Corixidae). Develop. Biol. 29: 330336. Scudder, G. G. E., M. S. Jarial and J. Choy. 1972. Osmotic and ionic balance in two species of Cenocorixa (Hemiptera). J. Insect. Physiol. 18: 883 895. -  Smith, B. P. 1977. Water mite parasitism of water boatmen (Hemiptera: Corixidae) M. Sc. Thesis Univ. BC Vancouver, Canada. 117pp. Smith, B. P. 1986. New species of Eylais (Acari: Hydrachnellae; Eylaidae) parasitic on water boatmen (Insecta: Hemiptera; Corixidae), and a key to North American larvae of the subgenus Syneylais. Can. J. Zool. 64: 23632369.  131 Smith, B. P. 1987. New species of Hydrachna (Acari: Hydrachnidia; Hydrachnidae) parasitic on water boatmen (Insecta: Hemiptera; Corixidae). Can. J. Zool. 65: 26302639. Smith, B. P. 1988. Host-parasite interaction and impact of larval water mites on insects. Ann. Rev. Entomol. 33: 487507. Smith, B. P. and S. B. McIver. 1984a. Factors influencing host selection and successful parasitism of Aedes spp. mosquitoes by Arrenurus spp. mites. Can. J. Zool. 62: 1114-1120. Smith, B. P. and S. B. McIver. 1984b. The patterns of mosquito emergemce (Diptera: Culicidae; Aedes. spp.): their influence on host selection by parasitic mites (Acari: Arrenuridae; Arrenurus spp.). Can. J. Zool. 62: 1106-1113. Smith, B. P. and S. B. McIver. 1984c. The impact of Arrenurus danbyensis Mullen (Acari: Prostigmata; Arrenuridae) on a population of Coquillettidia perturbans (Walker) (Diptera: Culicidae).^Can. J. Zool. 62: 1121-1134. Smith, B. P. and W. J. Cook. 1991. Negative covariance between larval Arrenurus sp. and Limnochares americana (Acari: Hydrachnidia) on male Leucorrhinia frigida (Odonata: Libellulidae) and its relationship to its host's age. Can. J. Zool. 69: 226-231. Smith, I. M. and D. R. Oliver. 1986. Review of parasitic associations of larval water mites (Acari: Parasitengona: Hydrachnida) with insect hosts. Can. Ent. 407-472. Spence, J. R. 1979. Microhabitat selection and regional coexistence in water-striders (Heteroptera: Gerridae). Ph.D. thesis, Universtity of British Columbia, Vancouver, Canada. pp 313. Stout, V. M. 1953. Eylais waikawae n. sp. (Hydracarina) and some features of its life history and anatomy. Trans. R. Soc. N. Z. 81: 389-416. Topping, M. S. and G. G. E. Scudder. 1977. Some physical and chemical features of saline lakes in central British Columbia. Syesis. 10: 145-166. Wiegert, R. G. and R. D. Mitchell. 1973. Ecology of yellowstone thermal effluent systems: intersects of bluegreen algae, grazing flies (Paracoenia, Ephydridae) and water mites (Partuniella, Hydrachnellae). Hydrobiologia 41: 251271.  132 Young, E. C. 1961. Degeneration of flight-musculature in the Corixidae and Notonectidae. Nature, Lond. 189: 328-329. Young, E. C. 1965a. Teneral development in British Corixidae. Proc. Roy. Entomol. Soc. London Ser. A. Gen. Entomol. 40: 159-168. Young, E. C. 1965b. Flight muscle polymorphism in British Corixidae: ecological observations. J. Anim. Ecol. 34: 353-389.  133 Appendix 1.  A. Life History Study of Eylais euryhalina Smith 1. Life cycle Figure 17 shows the life cycle of E. euryhalina throughout the summer at Becher's Prairie based on field collections and observations in 1990 and 1991. Dates are only approximations and some information is not known because the study was only carried out from May to September.- Letters below correspond to points on the life cycle diagram. A.  (May 4th)^Mites were almost entirely in parasitic  stages (larva or nymphochrysalis) although rare free-living nymphs were collected (1990). Larvae were either completely unengorged (and looked dessicated) or were approaching full engorgement. Nymphochrysalids displayed a wide range of development from freshly eclosed to showing full leg segmentation and ocelli development. B.  (June 2nd) Many free-living nymphs of E. euryhalina  were present in Long Lake and Lake Lye. ^Eggs of E. euryhalina were found on Scirpus spp. in Lake Lye (1991) (12 427 pS cm'), but did not hatch after several weeks in the lab (I suspect that they may have been from the previous summer). C. (June 9th) First E. euryhalina adult (female) was collected (1991) from Long Lake (13 848 pS cm -1 ), and  Figure 17. Life cycle of E. euryhalina at Becher's Prairie compared to the life cycle of Cenocorixa spp. Letters correspond to approximate dates as described in text pp. 133 - 136.  135 identified as such by the presence of an ovipositor (not visible in nymphs). D.  (June 19th) Many adult E. euryhalina were seen in  lakes. More eggs from Long Lake were collected, but they again do not hatch in the lab (1991). I was unable to find teleiochrysalids in field. E.  (June 27th) The first presence of Eylais spp. free-  living larvae on the surface of Lake greer (5 823 pS cm') was observed (1991). In the laboratory, they were identified as both E. euryhalina and E. discreta.^Egg masses of both species were found on Ruppia sp. and Zanischiella sp. floating vegetation and successfully incubated in the lab. F. (July 6th) Free-living E. euryhalina larvae were seen on Long Lake (1991). G.  (August 1st - 5th)^Freshly laid E. euryhalina egg  masses were no longer found in Lake Greer (E. discreta persisted for a week or two later), suggesting death of E. euryhalina adults that had over-wintered as parasitic larvae.  Free-living larvae were relatively scarce in Long Lake compared to mid-July.^A sentinel^study^(putting unparasitized hosts in a lake with parasites) in Long Lake showed only 10.3 % parasitism on C. expleta non-flying (N 29) and 0 % on C. bifida non-flying (N = 11) (August 5th to 12th, 1991).  136 H.  (September 13)^Second summer generation adult and  free-living E. euryhalina were present in Near Opposite Crescent Pond (4 261 pS cm -1 ) as well as Lake Lye. I. (October 21) Eggs collected from Lake Lye still hatch after incubation at 20 ° C for 24 hours. ^Adults cannot be found in Lake Lye or Long Lake. 2. Salinity range The presence of free-living larvae was used to determine whether E. euryhalina was breeding in a certain body of water. As shown in Table 1 (p. 13), E. euryhalina was found breeding from low salinity (Box 27: < 300 pS cm -1 ), through moderate salinity (Near Opposite Crescent Pond, Lake Greer, Near Pothole Lake, Long lake (Chilcotin), and Lake Lye (9 807 to 13 493 pS -1 ). Free-living larvae were not found in Barkley Lake (784 to 942 pS cm -1 ), but extensive collections were not taken. Field collections and observations of other life stages of E. euryhalina were also recorded. Engorging larvae and nymphochrysalids were found on flying hosts in all lakes including high salinity lakes such as Barnes, Round-up, and even LB2 (20 639 to 22 724 pS cm - '). Free-living nymphs were also recorded in small numbers from these lakes, but adults and egg masses were never found.  137 3. Mite engorgement process Following Lanciani (1971) and Reilly and McCarthy (1991), increase in area of the dorsal region was used to measure mite engorgement by the formula for area of an ellipse = 1/2 V (base X height).^The unengorged mites were 0.024 mm 2 (0.15mm long by 0.1 mm wide). ^Measurable engorgement was first witnessed 3.5 days after initial infection (2.5 days after hosts were removed from mite-infected water) with a slight increase in width. This was followed by an increase in length as the intersegmental membranes swelled beyond the posterior end of the dorsal shield. After engorgement began at 3.5 days, the mites' legs were entirely immobile, and by 8 days some of the mites had already attained a roughly spherical shape, unless in a site where full engorgement was inhibited. Some mites, however, remained unengorged after 6 to 8 days and closer examination revealed that they were dry and almost certainly dead. Under some of the unengorged mites were black, necrotic spots similar to that described for Hydrachna conjecta on Sigara falleni (Davids, 1973). The necrotic spot usually, but not always correlated with the unengorged state of the associated mite. Necrotic spots were not found on laboratory infected flying hosts as the reaction did not occur on the T.A.M. which was the only place where mites were able to commence engorgement. After 6 days, the  138 average size of mites for 2 replicates was 0.037 mm 2 (n = 49) and 0.038 mm 2 (n = 14) on C. bifida flying compared to 0.037 mm 2 (n = 4) and 0.025 mm 2 (n = 7) on C. expleta non-flying. After 8 days, C. bifida flying had mites of an average size of 0.031 mm 2 (n=2) and 0.045 mm 2 (n= 12) and C. expleta nonflying had mites of 0.047 mm 2 (n = 7) and 0.033  mm2  (n= 4).  Once again, these data were not statistically tested and are presented only to show that while engorgement was occuring, after 8 days the percentage of the total engorgement was very small. (A fully engorged E. euryhalina is nearly 2mm in diameter which equals 6.28 mm 2 .)  B. Discussion of E. euryhalina Life History 1. Life cycle From observations in the field and laboratory, some important facts about E. euryhalina have been discovered. Eggs are present all year, but do not remain viable throughout the winter as evidenced by early collections of eggs which appeared unembryonated and failed to hatch in the laboratory. I found degradation of overwintered eggs and prelarvae, such that larval features evident in the fall (i.e. eyespots and vitelline) were not apparent in the spring. Lanciani (1970a) states that only 3 of 20 Eylais spp. studied in North America overwinter exclusively in egg  139 stage, and E. euryhalina does not appear to be in this category. From early spring observations, E. euryhalina overwinters exclusively as a parasitic larvae. The potential infection period by E. euryhalina is established at 4 months: from late June until the lakes freeze in October with a decrease (or absence) of free-living larvae in August and September; the period of non-infectivity being dependent on the lake. It appeared that the less saline, usually smaller lakes (Near Opposite Crescent and Lake Greer) had earlier appearance of free-living larvae than the larger, moderate salinity lakes (Long Lake and Lake Lye). Shallow surface temperatures in these lakes are similar and show a small degree of variation between years (Scudder, 1975), so differences in mite life cycles may have been related to salinity, with lower salinity being quicker. (Free-living larvae were witnessed on June 27th in Lake Greer compared to July 6th for Long Lake.) Correspondingly, the absence of first generation free-living larvae in summer also occurred earlier in the smaller, less saline lakes as well. The actual dates of the appearance of the second generation free-living larval mites was not determined. By September 13, 1991, second generation free-living larvae were present on both Near Opposite Crescent Pond and Lake Lye.  140 From the presence of new, viable eggs in the field, it can be estimated that E. euryhalina laid eggs for about 35 days in the summer of 1991. With respect to the duration of the parasitic stage, an approximation of the period of overwintering for E. euryhalina can be judged by determining the parasitic duration of the latest possible attaching mite in the fall. Ice formed on the lakes at the end of October 1990, and assuming that the last attaching mite is the last one to leave the host in the spring (the end of May), the maximum possible time of the parasitic interaction (larvae and nymphochrysalis) is 7 months at this latitude (52 ° N). Lanciani (1969) states that larvae of the genus Eylais may be parasitic for a maximum of 11 months. The duration of the parasitic phase in the summer was not definitively determined, as laboratory rearing of mites on corixids in this study and others (Davids, 1973) has proven difficult. 2. Salinity range It was found that the upper salinity limit of E. euryhalina was not based on failure of the larval form to be  able to recruit or attach at high salinities. Experiment 6 demonstrated ability of the larvae to attach at salinities above the natural range of the species (albeit at low rates). Collections of nymphs in high salinity lakes suggests that inability to engorge is not the limiting factor, although  141 perhaps commencement of engorgement must occur at lower salinities before migration of the host to high salinity. The absence of adult mites in collections at these same high salinities implicates the development from nymph to teleiochrysalis, or teleiochrysalis to adult as the limiting life stage for E. euryhalina at high salinity. 3. Mite engorgement Because of the short duration of laboratory induced mite engorgement, no conclusions are drawn regarding the effects of host type on mite growth rate.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086201/manifest

Comment

Related Items