.ur 4w - x . “35:.“ ‘7‘. -A... :9 ' . d:“‘1 :L k _ . ‘ G“. .‘m - LG .21. ‘ . .u .2. ‘. ll‘xl .433” .x‘ijfi‘w". u‘ u. :1. t ”41 mm. .«M v. «.1? w: n . ‘1. ' “u“, \_ L l ..« k 1-3141 e . . _ . . 1 ‘ LL” 414-5” 253134“ - u :. 15.12..“1‘“ ( “h .. ' an ~ ‘\ «UN-“4.. . . ‘ L2 .~ » ‘ ' W ““"‘ 1;“13 F ‘ J“ k ~. Mr H “A...“ur. {i m.‘ m, ,.,.._,..’ ‘."‘ 1‘53 .-.4 1 . i' ' My. u .z.. . < . H. .. w up! .W «"1755. ~7th .. r .“u .u r I l \ Y . a ‘q I' i v _‘ . J ‘ 'n' 4-)] n 1"?“ m .44. "13-33553 I I ‘ A'I n V , . . A n. a. wfit wr- ‘ , _ . Ma ‘ A .‘ ‘ ,. WM . . x J ‘ ‘ru‘ r-‘ ' ' I V I _ . . :' " ._ 1. . I u ., '4 1 ‘ FHt‘sm This is to certify that the dissertation entitled Food partitioning in Collembola, with its relation to mouthpart structures and its effects on life history. presented by Benrong Chen has been accepted towards fulfillment of the requirements for Doctoral degree in Zoology ajor professor J. Snider, Ph.D. Date 54/7/7/ MSU is an Affirmative Action/Equal Opportunity Institution 0- 12771 lllllilillHIIIIWIilWlllllHllilllllHHHHIWIIIIIINI 31293 01025 9004 r LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. _ ‘— — 4— DATE DUE DATE DUE DATE DUE i o i i i MSU Is An Affirmative Action/Equal Opportunity lnflitutlon mundane-9.1 FOOD PARTITIONING IN COLLEMBOLA, WITH ITS RELATION TO MOUTHPART STRUCTURES AND ITS EFFECTS ON LIFE HISTORY by Benrong Chen A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1994 ABSTRACT FOOD PARTITIONING IN COLLEMBOLA, WITH ITS RELATION TO MOUTHPART STRUCTURES AND ITS EFFECTS ON LIFE HISTORY By Benrong Chen The present study was designed to assess mechanisms of food partitioning in Collembola. Isotoma notabilis, Sminthurinus henshawi and Orchesella harfasciata were dominant litter Collembola in deciduous forests in northern Michigan. Using scanning electron microscopy, mandibular and maxillary structures were described. Mandible and labrum were measured. Little overlap in mouthpart size occurred between species. Isotoma notabilis had the shortest mandible and labrum, 0. herfasciata the longest, whereas S. henshawz’ intermediate. Gut content analysis showed that these species fed on different types of food and selected food particles of different size. Selection of food particle size was related to the size of their mouthparts. Gut contents were categorized as fungal hyphae and spores, plant material and pollen, animal remains, colloidal materials, mineral matter and bacteria. Fungal hyphae _ 1". __ r“— _ . . _ and colloidal material were most abundant in the gut of I. notabilis, while S. henshawi fed heavily on fungal spores and 0. hexfasciata incorporated diversified foods in its diet. Type and size of food ingested were found to be species-specific. All species showed variation in dietary components related to seasonal availability of foods in the field. Four species of soil Collembola, six soil fungi and one actinomycete were isolated from soil, and were used in food preference tests in all possible combinations. Onychiurus armatus and Tullbergia granulata were selective feeders, showing very strong food preferences. Tullbergia yosiii and Tullbergia iowensis were general feeders, not discriminating between foods in most tests. One species in each of these groups was chosen for detailed life cycle studies in response to microbial foods. Diet (preferred ys. non-preferred foods) was found to affect 0. armatus significantly. Preferred foods generally resulted in higher egg production and survival. However, shorter lifespans were compensated for by higher survival and egg production during early life. In the general feeder T. yosiii, few effects on life history were seen. Results of this study suggest efficient partitioning of food sources among collembolan species coexisting in forest litter and soil, and document mechanisms for reducing interspecific competition. In memory of my father 7' ‘L‘n'l" 'wwl- lkk.‘ ACKNOWLEDGMENTS I wish to thank the members of my guidance committee: Dr. Thomas Burton, Dr. James Atkinson, Department of Zoology, Dr. John Lockwood, Department of Botany and Plant Pathology, and especially Dr. Renate Snider and Dr. Richard Snider, Chairman, for financial support and their aid in the preparation of this dissertation. I would also like to thank Dr. John Gill, Department of Animal Science, for help in statistical analyses, Dr. David Jacobson and Dr. Gerard Adams, Department of Botany and Plant Pathology, for help in microbial identification. Partial support for this study was provided by the Naval Electronic Systems Command through a subcontract to IIT Research Institute under contract NOOO39-81—C- 0357. TABLE OF CONTENTS Page LIST OF TABLES ...................................... ix LIST OF FIGURES ...................................... xi OVERVIEW .......................................... 1 CHAPTER I GUT CONTENT ANALYSIS AND MOUTHPART DESCRIPTION OF SOME COLLEMBOLA FROM NORTHERN MICHIGAN DECIDUOUS FOREST . . . 3 Introduction ...................................... 4 Site Description .................................... 10 Materials and Methods ............................... 12 Source of Specimens ............................ 12 Selection of Species ............................. 12 Sample Preparation and Analyses ..................... 14 Results ......................................... 15 Mouthparts .................................. 15 Mandibles .............................. l6 Maxillae ............................... 29 Labrum ................................ 31 Number of Individuals with Gut Contents ................ 34 Food Particle Size Distribution ...................... 42 vi Food Components .............................. 49 Discussion ....................................... 64 CHAPTER II FOOD PREFERENCE AND EFFECTS OF FOOD TYPE ON THE LIFE HISTORY OF SOME SOIL COLLEMBOLA ............... 72 Introduction ...................................... 73 Materials and Methods ............................... 78 Species ..................................... 78 Food Preference Tests ........................... 80 Life History Studies ............................. 81 Results ......................................... 83 A. Food Preference ............................. 83 B. Growth ................................... 86 Onychiurus afinatus ........................ 86 T ullbergia yosiii ........................... 88 C. Survival .................................. 88 Onychiurus armatus ........................ 88 Tullbergia yosiii ........................... 90 D. Egg Production ............................. 90 Onychiurus armatus ........................ 9O Tullbergia yosiii ........................... 93 E. Production of Exuviae by T. yosiii .................. 98 Discussion ....................................... 98 SUMMARY ......................................... 105 vii IMPROVEMENT AND FURTHER STUDY ..................... LIST OF REFERENCES ................................. Appendix 1. Multiple regression for estimating the width of the labrum of each species. .................................. Appendix 11. Weekly body length (um, means i SD) of 0. armatus fed on three different microbial foods. ......................... Appendix III. Weekly body length (um, means -_|; SD) of T. yosiiz' fed on three different microbial foods. ......................... Appendix IV. Weekly survival rate (means 1- SD) of 0. armatus fed on three different microbial foods. ......................... Appendix V. Weekly survival rate (means i SD) of T. yosiii fed on three different microbial foods. ......................... Appendix VI. Weekly cumulative egg production (means i SD) of 0. armatus fed on three different microbial foods. .................. Appendix VII. Weekly cumulative egg production (means _-|; SD) of T. yosiii fed on three different microbial foods. .................... Appendix VIII. Weekly cumulative exuviae production (means i SD) of T. yosiii fed on three different microbial foods. .................. viii 109 111 119 120 122 124 127 130 133 136 LIST OF TABLES Table Page 1. Average percent of individuals with gut contents. ................. 36 2. Analysis of variance of the percent of individuals with gut contents in 1988. 10. 11. Tested are effects of Species (I. notabilis, S. henshawz' and 0. hexfasciata), Site (I and II), Week and their interactions. .......... 37 . Analysis of variance of the percent of individuals with gut contents in 1988. Data from Site I and H are combined. ...................... 38 . Analysis of variance of the percent of 0. haxfasciata with gut contents. Tested are Year (1988 and 1990), Site (I and H), Week and their interactions. ...................................... 40 . Analysis of variance of the percent of 0.hexfasciata with gut contents in 1990. Tested are Site, Week and their interaction. .............. 41 . Analysis of variance of the percent of 0.hexfasciata with gut contents in each site. Tested are Year (1988 and 1990), Week and their interaction. . . . . 43 . Multivariate analysis of variance of food particle size distribution in 1988. Tested are Species, Site, Week and their interactions .............. 50 . Multivariate analysis of variance of food particle size distribution in 0. hafasciata. Tested are Year, Site, Week and their interactions. ...... 50 . Multivariate analysis of variance of food particle size distribution in each species and year. Tested are Site, Week and their interaction. ....... 51 Percent (means i standard errors) of dietary components in the guts of three collembolans. (N = total numbers of specimens examined). ........ 52 P-values of analysis of variance of food items in 1988. Tested are Species, Site, Week and their interactions. ......................... 54 ix 12. Relative percentage (based on the total amount of pollen) of pine and non- pine pollen found in the guts of S. henshawi and 0. hexfasciata. ...... 55 13. P-values of analysis of variance of food items of 0. hexfasciata. Tested are Year, Site, Week and their interactions. ..................... 57 14. P-values of analysis of variance of food items of individual species. Tested are Site, Week and their interaction. NP = none present. ......... 58 15. Summaries of results of feeding preference experiments. ............. 84 LIST OF FIGURES Figure Page 1. Mandibular structure of 0. hexfasciata. A. Left mandible, ventral view. 260 x. B. Right mandibular molar plate, median view. 1,300 x. C. Molar plate buttress rods. 9,400 x. D. Dorsal edge of molar plate. 7,200 x. . . 17 E" Mandibular and maxillar structure of 0. harfasciata. A. Left mandibular molar plate end. 4,400 x. B. Right mandibular molar plate end. 4,000 x. C. Left maxillar head, dorsal view. 1,200 x. D. Left maxillar head, mid- ventral view. 1,300 x. E. Left maxillar head, median view. 1,300 x. F. Right maxillar head, ventral view. 1,800 x. .................. 19 3. Mandibular structure of S. henshawi. A. Dental portion of right mandible, ventral view. 1,000 x. B. Dental portion of left mandible, median view. 1,100 x. C. Molar plate buttress rods, ventral view. 20,000 x. D. Dorsal edge of left molar plate. 5,400 x. .................... 21 4. Mandibular and maxillar structure of S. henshawi. A. Right mandibular molar plate end. 9,400 x. B. Left mandibular molar plate end. 9,400 x. C. Left maxillar head, ventral view. 3,000 x. D. Left maxillar head, median view. 3,300 x. ............................... 23 5. Mandibular and maxillar structure of I. notabilis. A. Dental portion of left mandible, median view. 2,400 x. B. Right mandibular molar plate end. 13,000 x. C. Left mandibular molar plate, median view. 4,800 x. D. Right mandibular molar plate end. 7,800 x. E. Left mandibular molar plate end. 12,000 x. F. Left maxillar head, median view. 3,600 x. . . . . 25 6. Mandible length frequency distribution of three collembolan species. ..... 28 7. Labrum of three collembolan species. A. Isotoma notabilis. B. Sminthurinus henshawi. C. Orchesella hexfasciata. .............. 32 8. Labrum width frequency distribution of three collembolan species. ...... 35 xi . Percent (means -l_- standard errors) of individuals with gut contents for the three species from each site. A: I. notabilis; B: S. henshawi; C: 0. herfasciata 1988 and D: 0.hexfasciata 1990. Filled bar: Site 1; Empty bar: Site 11. ...................................... 39 . Average food particle size (means i standard errors) distribution of three collembolan species. Small: < 6 umz. Medium: 6 to 20 umz. Large: > 20 um2. ...................................... 44 Seasonal food particle size (means i standard errors) distribution in I. notabilis. ........................................ 45 Seasonal food particle size (means i standard errors) distribution in S. henshawi. ....................................... 46 Seasonal food particle size (means i standard errors) distribution in 0. hexfasciata, 1988. .................................. 47 Seasonal food particle size (means i standard errors) distribution in 0. hexfasciata, 1990. .................................. 48 . Seasonal food components of I. notabilis. A: Site I. B: Site 11. ....... 60 . Seasonal food components of S. henshawi. A: Site I. B: Site 11. ....... 61 17. 18. 19. 20. 21. 22. 23. Seasonal food components of 0. haxfasciata, 1988. A: Site I. B: Site 11 . . 62 Seasonal food components of 0. hexfasciata, 1990. A: Site I. B: Site H . . 63 Body length of 0. armatus fed on three different microbial foods. Vertical lines represent standard deviations. ........................ 87 Body length of T. yosiii fed on three different microbial foods. Vertical lines represent standard deviations. ........................ 89 Mean survival rates of 0. armatus fed on three different microbial foods. Vertical lines represent standard deviations. ................... 91 Mean survival rates of T. yosiii fed on three different microbial foods. Vertical lines represent standard deviations. ................... 92 Mean weekly egg production per individual 0. armatus fed on three different microbial foods. .............................. 94 xii 26. 25. 26. 27. Mean cumulative egg production per individual 0. armatus fed on three different microbial foods. Vertical lines represent standard deviations. . . mean weekly egg production per individual T. yosiii fed on three different microbial foods. ................................... Mean cumulative egg production per individual T. yosiii fed on three different microbial food. Vertical lines represent standard deviations. Mean cumulative exuviae production per individual T. yosiii fed on three different microbial foods. Vertical lines represent standard deviations. . . xiii 95 99 OVERVIEW Collembola (Insecta) are one of the largest arthropod populations living in litter and soil. Their densities range from 100 individuals m”2 in a desert (Wallwork, 1972) to 670000 m'2 measured in permanently moist, ornithogenic soils (Collins et al., 1975). '2 in temperate deciduous forests Mean annual densities range between 40 and 70 x 103 m (Petersen, 1982). Although the amount of soil metabolism that can be attributed to soil Collembola is relatively low, the concept that soil Collembola regulate the decomposition process has become more and more popular (Moore et al, 1988). Fungi and bacteria are directly responsible for most organic matter breakdown. By greatly influencing decomposer flora, Collembola contribute significantly to decomposition and nutrient cycles (Seastedt, 1984). They have an integral role in maintaining and shaping microbial activity and community structure, and are important mediators of food—web stability (Moore et al, 1988). These functions are believed to be firmly associated with their feeding activities (Seastedt, 1984), and are affected dramatically by environmental conditions. The high density and species richness of Collembola in almost every habitat have directed a lot of attention to food partitioning mechanisms among coexisting species (Fjellberg, 1985). In early studies, especially from the late 50’s to the early 80’s, most 2 attention was paid to gut content analysis of field populations (Poole, 1959; Knight and Angel, 1967; Bodvarsson, 1970; Gilmore and Raffensperger, 1970; Takeda and Ichimura, 1983). In contradiction to the competitive exclusion theory, most results indicated that coexisting Collembola showed a very low degree of differentiation in food consumption (Anderson and Healey, 1972; Takeda and Ichimura, 1983). In recent years, emphasis has been shifted to laboratory studies, dealing especially with food preference and life histories, and results have documented that food preferences did exist in Collembola (Visser and Whittaker, 1977; Shaw, 1988; Schultz, 1991). Studies of collembolan food and feeding biology provide further useful information concerning the functions of these animals in soil ecosystems. On the other hand, feeding habits of Collembola also affect their population dynamics. Three general categories of methods for obtaining trophic information about soil animals are 1) direct observation; 2) gut content analysis; 3) experiment/inference (Walter et al, 1991). CHAPTER I GUT CONTENT ANALYSIS AND MOUTHPART DESCRIPTION OF SOME COLLEMBOLA FROM NORTHERN MICHIGAN DECIDUOUS FOREST Introduction MacNamara’s paper (1924) was the first important work on the feeding habits of Collembola. He divided Collembola into two groups based on their mouthpart structures: chewing forms which have a well-developed mandibular molar plate, and suctorial forms which do not have a mandibular molar plate. Chewing forms were thought to be vegetarian and suctorial forms carnivorous. The majority of collembolan species are of the chewing type, and their numbers far exceed the suctorial type in most environments. Examining the gut contents of Collembola, MacNamara found them to consist of decaying plant material, fungal hyphae and spores, and minute algae in chewing forms, and suggested that fungi were a favorite food for these animals. Liquid food also attracted some of them: Achorutes = Hypogastrura spp. were found feeding on maple sap. Isotoma palustris Miiller = Isotomurus palustris Miiller and Sminthurz'des aquaticus (Bourlet) often picked up diatoms, desmids and conifer pollen in spring. He also observed Achorutes annatus Nicolet = Hypogastrura armata (Nicolet) feeding on mushrooms and Achorutes viaticus (Tullberg) = Hypogastrura viatica (Tullberg) on a colloidal sewage deposit. MacNamara (1924) was somewhat suspicious of the "grinding" function of the mandibular molar plate. Since there were often fair-sized particles of wood and uncrushed fragments of hyphae and spores in the guts, he thought these animals did not always use the mandibular molar plate to grind food, but swallowed much of their food whole. It was concluded by Agrell (1940) that soil Collembola were unspecialized 5 feeders. The gut contents of Collembola in one locality were filled with fungal hyphae, while in another locality the same species had guts full of amorphous detritus. Generally Collembola do not feed on fresh leaves. Tomocerus flavescens (Tullberg) and Orchesella flavescens (Bourlet) would feed on decaying leaf litter and T. flavescens flourished on this diet (Schaller, 1949). When excrements of larger arthropods, e.g., Glomeris sp., Julus sp. and Porcellium sp. were offered to some collembolan species, T. flavescens and Onychiurus armatus (Tullberg) would eat them but Isotoma olivacea Tullberg would not. Schaller also examined collembolan gut contents and found that fungal hyphae were the dominant constituent, with decaying wood in some species. He concluded that Collembola may be important in comminution of litter and humus. Murphy and Doncaster (1957) found that Onychiurus armatus fed on nematodes of the genus Heterodera sp.. But their results cannot be regarded as really conclusive since many Collembola eat almost anything offered to them. Poole (1959) studied gut contents of several collembolan species from a single habitat (Douglas fir plantation). He observed that about half of the individuals of each species of chewing type had no visible gut contents; neither did most of the suctorial type species. Fungal hyphae and spores made up the majority of recognizable material, higher plant tissue was present in small amounts and mineral particles were common. Amorphous debris varied considerably in amount, but generally represented a larger proportion in smaller animals. Other materials found in the gut were animal remains, including amoebae, earthworm setae and collembolan scales. Quantitative estimates of selected gut contents (fungi, plant materials and mineral particles) were done for several l—M 6 coexisting species. Results showed that large species (such as Tomocerus minor (Lubbock) and Tomocerus longicorm's (Miiller)) fed mainly on soil fungi, whereas smaller forms appeared to feed directly on humus. There were no marked differences in feeding habits among chewing type species; but there was some evidence of seasonal variations. Poole (1959) related the seasonal changes to either a scarcity of soil fungi or an abundance of palatable litter remains. Viable soil fungi passed through the gut of T. longicomis indicated that Collembola probably play an important role in the dispersal of soil fungi. In a review paper, Christiansen (1964) pointed out that there were very few carnivorous soil Collembola, and that most species appeared to use a wide variety of foods, including fungi, plant material and bacteria. He stated that the diet of an animal is essentially what is available. Contrary to Poole’s (1959) results, Knight and Angel (1967) found that Tomocerus flavescens fed more heavily on leaf material than on fungi in a hardwood community. They attributed the difference to species and/ or habitat differences. They also found that when offered different foods, the species selected fungal spores far more often than either leaf material or fungal hyphae. It was speculated that discrepancies in spore content between laboratory and field animals were probably a result of the distribution and density of spores in the field. Attempting to determine whether variations in gut contents were correlated with habitat dissimilarities or were species—specific (questions raised by the work of Poole (1959) and Knight and Angel (1967)), Gilmore and Raffensperger (1970) found that variations in gut contents of Tomocerus spp. were related to habitat characteristics. In 7 addition, their data did not show evidence to support food preferences between various Tomocerus species. They pointed out that preferences for fungi or for certain kinds of leaves may exist, but that they were undetectable by the methods employed. It was further suggested that various species of Tomocerus could occupy different micro- habitats, while at the same time feeding upon nearly identical materials. In part, results on interspecific differences were obscured by small sample sizes combined with high variation between individuals of a given species. Some studies (Thibaud, 1968; Joosse and Veltkamp, 1970; de With and Joosse 1971) have indicated that empty guts in Collembola were associated with molting. At any one time only 50-60% of the animals were feeding, and molting animals were Weeding. Anderson and Healey (1972) suggested that empty guts in Collembola probably indicated lack of feeding rather than feeding on indistinguishable food substances. They measured the proportion of gut particles larger than 10 um in three entomobryids from one habitat. Their work was the first to show seasonal variation in collembolan feeding habits. They found some degree of difference in gut contents among these three species. Plant material in Orchesella flavescens appeared to be in a less advanced stage of decomposition than that in two Tomocerus species, and T. minor usually took more fungal food than T. longicornis when both were present in the samples. They concluded that, compared to invertebrate herbivores, there was generally less evidence for food differentiation in Collembola. Three hypotheses were proposed to explain this phenomenon: 1) there is an excess of food available to decomposers; 2) there may exist undetected differences between species in their utilization of resources; 3) there are 8 undetected micro-habitat differences between species. Their preliminary results did not show any relationship between food particle size and body size of two sympatric species of Tomocerus. They concluded that if any selection of food particles by size occurred, it was likely that differences were lost during the grinding or digestion processes. In a study of gut contents of sympatric onychiurids, McMillan (1975) also found a low degree of differentiation in food consumed. His data suggested, however, that some particle size selection occurred: the largest species, Onychiurus fiircifer (Borner), contained a higher proportion of thicker fungal hyphae than the other two onychiurids. Significant food differentiation was observed among three coexisting entomobryids studied by Muraleedharan and Prabhoo (1978). In addition to high percentages of an amorphous substance (a mixture of dietary components which had undergone partial digestion) in the guts of both Alloscopus tetracantha Borner and Cyphoderopsis decemoculata Prabhoo, A. tetracantha consumed more higher plant material and very little fungal material; on the other hand, C. decemoculata preferred fungal spores to higher plant material. Microparonella duodecemoculata Prabhoo had a high proportion of fungal spores in their guts, and spores found in this species were totally different from those in C. decemoculata. The authors also suggested that the small percentage of higher plant material in the gut of M. duodecemoculata and C. decemoculata, and fungal material in the gut of A. tetracantha, may have been due to accidental intake rather than to deliberate feeding. In contrast to the investigation of Muraleedharan and Prabhoo (1978), Takeda and Ichimura (1983) concluded that even species inhabiting distinctly different habitats can have the same food preferences in the field. Their results were based on gut content 9 analysis of two litter- and two soil-dwelling species from a pine forest. There was a high degree of overlap in food items recorded for all four species. Collembola with a mandibular molar plate use their mouthparts (mainly mandibles and maxillae) to chew, grind and/or tear food material prior to ingestion. The tips of the mandibles can scratch the food surface, and mandibles and maxillae can also reach out, grasp and carry food back into the mouth. Movement of mandibles and maxillae is small because of restriction by surrounding structures, such as labrum, labium and the lateral margins of the gnathal pouch (Manton, 1964). Shape and size of mandibles, maxillae and labrum may restrict type and particle size of food ingested by Collembola (F j ellberg, 1985), and therefore may contribute to niche separation of Collembola through their feeding habits. Although the importance of mouthpart structures in feeding habits has long been recognized (MacNamara, 1924; Fjellberg, 1985), no attempt has been made to investigate the relationship between detailed mouthpart structures and food selection in Collembola. Although built around the general same model, the actual anatomy of individual parts of the feeding apparatus varies considerably, even in closely related systematic groups (Fjellberg, 1984). Cummins (1973) has suggested that food may be partitioned largely on a basis of particle size. Obviously, size of individual particles consumed could be as important in niche segregation as type of particles ingested. Most gut content analyses concentrated on types of collembolan foods; very few studies mentioned food particle size differentiation among species (Anderson and Healey, 1972; McMillan, 1975). Comprehensive studies of food particle size in Collembola have not been reported. According to the competitive exclusion theory, coexisting species must have some 10 mechanisms for separation. Presumably, coexisting collembolan species should show food selectivity, and/ or be separated on the basis of micro-habitat or of behavioral characteristics. If food selectivity exists, it should be reflected in their gut contents (food type and particle size). The major hypothesis to be tested in the present study is that there is some differentiation in mouthpart structure and size among coexisting Collembola; and that these differences will affect the feeding habits of these species (as reflected in the food types and particle sizes selected). Other hypotheses are that seasonal variation in gut contents reflects the availability of food in the field; and that animals living under similar environmental conditions have similar gut contents. The present study therefore assessed the following aspects of the ecology of some coexisting collembolan species: a) Description and measurement of mouthpart structures (mandibles, maxillae and labrum); b) Analysis of species—specific gut contents, and their variability with respect to seasons, sites, and years; c) Relationship between mouthpart structures and gut contents; (1) Effects of seasonal and environmental conditions on the gut contents. Site Description Collembola observed in this study came from the Test and Control sites used for soil biological studies (ELF ecological monitoring project) in 1988 and 1990. These two sites were closely matched with respect to physical and chemical characteristics (Snider 1 l and Snider, 1987). The "Test" site of Project ELF will be referred to as Site I and the " Control" as Site H in the present study. Both sites are located in Dickinson County, Upper Peninsula, Michigan, and are separated by a distance of 11.5 km. Detailed descriptions of the study areas are given in Snider and Snider (1987). The climate of the general area is temperate continental of the cool summer type. Annual mean precipitation is 756 mm, evenly distributed, and snowfall occurs from September to May. Yearly average temperature is 5.40C. Air temperatures near the forest floor were virtually the same in both sites, daily means generally differing by less than 1°C. The two sites were secondary growth deciduous forests. Both were dominated by maple (Acer saccharum March), with basswood (Tilia americana L.) subdominant. Among minor stand elements, Osttya virginiana (Mill) K. Koch and Populus spp were equally common in both sites’ understory. The shrub association was dominated by leatherwood (Dirca palustris L.), with a higher density in Site H. Sedge species were very abundant in both sites as ground cover. Pine trees were located in the adjacent areas of both sites. There was no evidence of recent physical disturbance in either site. Annual leaf litter inputs in both sites were approximately equal. Leaf abscission generally peaked in early October. Soils in both sites have developed on coarse- to medium-textured glacial till and are classified as naturally well-drained Spodosols (podzols), with the A horizon tending strongly toward mull. Soil texture was very similar, with sand predominating (about 60%) in both sites. 12 Materials and Methods Source of Specimens Site I and II were each divided into 20 (10 x 10 m) quadrats, from which samples were obtained periodically. Collembola used for gut content analysis stemmed from two types of samples: a.) Pitfall traps: Each quadrat contained a permanently installed pittrap, with ethylene glycol as the collecting medium. After samples were gathered, they were washed into 95% ethanol until further treatment. b.) Leaf litter Samples: Litter samples (25 x 25 cm) were randomly collected from ten quadrats on the forest floor at intervals of two weeks. The samples were then placed in Tullgren funnels for about eight days to extract animals. The heat in the funnels was low at first, and was gradually increased every day. Animals were collected in jars containing 95% ethanol. Selection of Species Three species were chosen for the experiment: Isotoma (Deson'a) notabilis Schaeffer (extracted from leaf litter), and Sminthurinus (Sminthurinus) henshawi (Folsom) and Orchesella hexfasciata Harvey (both obtained from pittraps). They were selected because of their relatively high dominance in both sites, and because of their consistent occurrence from May to October. Isotoma notabilis represented 56% of all Collembola obtained in litter samples from Site I and 62% of those obtained from Site 11. 13 Sminthurinus henshawi represented 30% and 0. hexfasciata 26% of all Collembola captured in pitfall traps in 1988, and 21% of the total in 1990 were 0. hexfasciata (Snider and Snider, 1990, 1992). For assessing the number of individuals with gut contents, all specimens (immature and adult) were used in the case of S. henshawi and 0. hexfasciata, but only adult I. notabilis were chosen because it was difficult to determine whether immatures contained any food at all. For gut content analysis and mouthpart description, only adult animals of all three species were used. Since they are ametabolous, adult Collembola can only be separated from immatures by the presence of a genital plate. A set of differently-sized specimens of each collembolan species was measured, and mounted on slides. These specimens were observed under a phase-contrast microscope for presence of a genital plate. Results showed that after reaching a certain body length (including head), all specimens were adults. For I. notabilis, S. henshawi or 0. herfasciata, body lengths greater than 540, 290 or 840 um, respectively, were determined to be adults. Interspecific, within-year comparisons were based on 1988 data for all three species. In the case of 0. hexfasciata, 1988 and 1990 data were used for between-year intraspecific analyses. Analyses of seasonal differences were based on samples obtained at intervals ranging from two to five weeks. Although pittrap samples were available at weekly, and litter samples at biweekly intervals, only samples with sufficient numbers of individuals from both sites were used here (for a total of seven or eight dates per year). In the subsequent presentation of results, "Week 1" through "Week 24 " will span a time period 14 from early May through mid-October. Sample Preparation and Analyses For each sample and species, the number of individuals with gut contents was counted using a stereomicroscope, counts being expressed as percent of total number. Percentages (p) were transformed by loge [p/ (1-p)]; in cases where 100% or 0% of animals had gut contents, the data were substituted by (1 — 0.25/n) or (0.25/n) (where n = total number of that species found in that sample) respectively. Statistical analyses were then performed using transformed data. Before dissection, head and trunk length of each individual were recorded. Mandibles and maxillae were separated from head capsules and mounted in a small drop of Hoyer’s medium (Borror et al, 1976) on a glass slide. Total length of the mandible was measured under a phase-contrast microscope. Mandibles and maxillae were examined for their detailed structure. Labra were at first discarded, but were later included in the study. Dissection of an additional set of specimens showed that labrum width could be reliably estimated by regression (R2 > 0.95) based on head, trunk and mandible length (Appendix I). Some mandibles and maxillae were dissected for scanning electron microscopy. Mouthparts were dissected in 100% ethanol, mounted on stubs using double stick-tape, air dried, sputter—coated with gold, and examined in a J EOL J SM—35CF scanning electron microscope. Gut contents of individual specimen were dissected and dispersed in Canada Balsam on a glass slide. They were examined under a phase—contrast microscope at 750x 15 magnification. Eight food categories were identified as follows: fungal hyphae, fungal spores, plant materials, pollen, animal remains, minerals, bacteria and colloidal material. A small portion of the gut content could not be identified based on current knowledge and was placed into an "unknown" category. Each food category was measured using a grid eye-piece and the data were converted into percentage for each specimen. Percentages were then transformed by square root to achieve normality and analyzed. Pollen was further separated into pine pollen and non-pine pollen. Relative percentages of pine and non-pine pollen were calculated based on the total pollen amount, and nonparametric statistics was used to analyze data. Detailed gut content records were based on a total of 858 specimens (204 specimens of I. notabilis, 234 S. henshawi, 203 0. hexfasciata in 1988 and 217 in 1990). For each individual, approximately 100 food particles, randomly chosen, were measured for their two-dimensional size and grouped into three (small, medium and 2 large) size classes. Small corresponded to particles smaller than 6 um , medium 2, and large ones were > 20 umz. The raw particles measured between 6 and 20 um data were converted to percentages (p) and then transformed by loge [p/(l-p)] to achieve normality before statistical analysis. Results Mouthparts The mouthparts of Collembola are prognathous and entognathous. They are contained in the cavity formed by the fusion of the lateral margins of the labium with the 16 head (Folsom, 1900; Manton,1977). The mandibles lie above (i.e., morphologically anterior to) the maxillae and dorsa-lateral to the hypopharynx (Goto, 1972). Mandibles The mandibles of these three collembolan species are all of the "biting type" (Wolter, 1963), i.e. they are provided with a well-developed molar plate in addition to the apical incisor teeth. The general shapes of the mandibles are very similar to those of Tomocerus longicomis described by Hoffmann (1905, 1908) and Manton (1977), and Folsomia candida by Goto (1972). They consist of two parts: an anterior or dental portion, and a posterior portion or fulcrum, for articulation and muscular insertion (Folsom, 1899) (Figure 1A). The apex of the mandible bears some sharp, incisive teeth termed apical teeth on its median side. The right mandible of 0. hexfasciata consistently has five teeth . Most individuals of I. notabilis have four teeth, a few have five. Sminthurinus henshawi usually has five teeth except for some individuals with six (Figure 3A). The left mandible of these three species invariably has four teeth. The whole region bearing apical teeth is bent slightly downwards (Figure 3B, 5A). The molar plate is crescentic shaped (Figure 3B, 5A). At a point about mid-way along its length, the plate as a whole turns upwards, especially in S. henshawi and I. notabilis (Figure 3B, 5A). Ventrally, the molar plate has conical cusps. Like apical teeth, they gradually decrease in size from distal to proximal. Immediately following the cusps are structures called "buttress rods" by Goto (1972). There are a few fmger-like apical processes at the tips of each of these buttress rods. These finger-like apical 17 Figure 1. Mandibular structure of 0. herfasciata. A. Left mandible, ventral view. 260 x. B. Right mandibular molar plate, median view. 1,300 x. C. Molar plate buttress rods. 9,400 x. D. Dorsal edge of molar plate. 7,200 x. d = dental portion. f = fulcrum. c = conical cusp. r = buttress rod. dorsal boundary rods. 1:: ll 18 Figure l. 19 Figure 2. Mandibular and maxillar structure of 0. hexfasciata. A. Left mandibular molar plate end. 4,400 x. B. Right mandibular molar plate end. 4,000 x. C. Left maxillar head, dorsal View. 1,200 x. D. Left maxillar head, mid-ventral view. 1,300 x. E. Left maxillar head, median view. 1,300 x. F. Right maxillar head, ventral view. 1,800 x. p = dorsal boundary rod. u = ungulum. 1 = lamella l. 2 = lamella 2. 4 = lamella 4. 5 = lamella 5. 6 = lamella 6. 20 21 Figure 3. Mandibular structure of S. henshawi. A. Dental portion of right mandible, ventral View. 1,000 x. B. Dental portion of left mandible, median view. 1,100 x. C. Molar plate buttress rods, ventral view. 20,000 x. D. Dorsal edge of left molar plate. 5,400 x. = apical teeth. 0 = conical cusp. buttress rod. .., II p = dorsal boundary rod. 22 Figure 3. 23 Figure 4. Mandibular and maxillar structure of S. henshawi. A. Right mandibular molar plate end. 9,400 x. B. Left mandibular molar plate end. 9,400 x. C. Left maxillar head, ventral View. 3,000 x. D. Left maxillar head, median view. 3,300 x. u — ungulum. 2 = lamella 2. 3 = lamella 3. 4 = lamella 4. 5 = lamella 5. 6 = lamella 6. 24 Figure 4. 25 Figure 5. Mandibular and maxillar structure of I. notabilis. A. Dental portion of left mandible, median View. 2,400 x. B. Right mandibular molar plate end. 13,000 x. C. Left mandibular molar plate, median view. 4,800 x. D. Right mandibular molar plate end. 7 ,800 x. E. Left mandibular molar plate end. 12,000 x. F. Left maxillar head, median view. 3,600 x. t = apical teeth. c = conical cusp. r = buttress rod. 1 = lamella 1. 2 = lamella 2. 3 = lamella 3. 4 = lamella 4. 5 = lamella 5. 6 = lamella 6. 26 Figure 5. 27 processes are fairly long in 0. hexfasciata (Figure 1C), but short in both S. henshawi (Figure 3C) and I. notabilis (Figure 5B). The main body of the molar plate consists of many regular transverse rows of multicuspid molar rods. Each of these rows consists of a row of higher, simple rods at equal intervals. These rods correspond to the "teet " or cusps under a light microscope. They are surrounded by some lower, multi-headed rods. In I. notabilis, these multicuspid molar rods are uniformly distributed from the distal to the proximal end of the molar plate (Figure 5C). In 0.he.wg”asciata, the proportion of simple rods and multi—headed rods decreases from the distal region to proximal, simple rods being absent from the ventral-proximal region (Figure 1B). In S. henshawi, some dorsal boundary rods have expanded heads (Figure 3D). At the extreme proximal end of the right molar plate, there is a very large, apically rounded projection that points inward towards the middle of the head (Figure 2B, 4A, 5B, 5D). On the left, there is a depression at that same position (Figure 2A, 4B, 5B). It could be that the right projection and left depression keep the two mandibles held together. The length of the mandible differs in these three collembolan species. Isotoma notabilis has the shortest, ranging from 81 to 130 um with more than 90% of samples shorter than 120 um. Mandible size of S. henshawi is intermediate, ranging from 92 to 216 um with more than 90% in the range of 120 to 200 um. Orchesella hexfasciata has the longest, in the range of 164 to 403 um with more than 95% longer than 200 um. There is very little interspecific mandible length overlap (Figure 6). The strong apical teeth of all three species have the ability to out food materials. The main molar plate surfaces in all three species appear flexible, and may not be very 28 4o ' l I ————— 0 hex/3.90313 30 _ ............. 8 ”608/739", / 00/80/719 g 3 20 - 1'- h a) I: o_ 3 10 ‘ . l’\\ _ fl VAL. \ / \ r \ A \ ’...J/ I \\ A\ o r ‘ " I ' d 50 150 250 350 450 Mandible Length (0) Figure 6. Mandible length frequency distribution of three collembolan species 29 effective for grinding as previously thought; rather, they may be of a brushing nature. Although some minor anatomical differences were observed between these three species, it is difficult to explain how these differences would affect feeding habits. On the other hand, mandible size differences suggest some differences in feeding ability (especially the size of food particles that they can handle) among species. Maxillae The elongate stipes of the maxillae articulate basally via the corda and corda- posterior tentorial membrane, with the transverse process of the posterior tentorial apodeme (Manton, 1964). The stipes extend forward to articulate with the outer lobe ("palp" of Folsom (1899), Hoffmann (1905, 1908) and Manton ( 1964)) and the maxillar head. The principal structures of the maxillar head of the three species studied here (Figures 2, 4, 5) are very similar to those of Folsomia candida as described by Goto (1972) and of Hypogastruridae (Fjellberg 1984). The maxillar head consists of a three- toothed ungulum and some lamellae. Fjellberg’s system is followed for lamellar naming. Isotoma notabilis is very similar to Folsomia candida in number and shape of lamellae (Figure 5F). Lamella 1 (median appendage of B0rner,1908 and Goto, 1972) projects well beyond the ungulum teeth and branches to form the apical rake consisting of three irregular rows of fringes. The distal row has the most fringes and proximal row the fewest. The shape of lamella 6 (median subsidiary appendage of Goto, 1972) differs from that of F. candida in having five irregular rows of fringes. Lamellae 4 and lamella 5 are located dorsally on the maxillar head and both have proximal fringes. Two 30 lamellae are situated on the ventral side, lamellae 2 and 3. Lamella 2 extends to the height of the ungular teeth and only has fringes on its proximal edge. Lamella 3 is a small one bearing fringes on its proximal side. There are only five lamellae on S. henshawi (Figure 4C, 4D). Lamellae 4 and 5 bear few proximal fringes. Two lamellae on the ventral side correspond to lamella 2 and 3. Lamella 2 is the largest of these five lamellae. It is well equipped with fringes on its proximal side. Lamella 3 is triangular and very small. There is one lamella in the middle, lamella 6. A very small projection at the base, because of its shape and position, is not thought to be a lamella, but its origin is unknown. In 0. hexfasciata there is a well developed ungulum with three strong teeth dorsal to its main axis (Figure 2C). These teeth curve medially (Figure 2E). Lamella l is barely longer than the ungular teeth, the basal shaft arises from the ungulum at very low level and curves ventrally. The apical portion is not branched to form a "rake" as in Folsomia candida, instead it gradually points distally, bearing fringes and ends at ventral side of the apical angular tooth (Figure 2D). Lamella 6 is a pad-shaped structure (Figure 2E) located in front of lamella 1. Its concavity and edge are well supplied with fringes. Lamellae 4 and 5 are located dorsally. These two lamellae are very much alike, bearing proximal fringes. They are fused basally with the proximal ungulum teeth (Figure 2C). There is only one ventral lamella, lamella 2. This lamella is stout, folded along its mid- line, and has fringes laterally and ventrally. It connects to lamella 6 dorsally and is separated ventrally from the ungulum teeth (Figure 2F). The maxillar head is complex, and the basic structures of these "biting type" collembolan species are very similar in having three ungular teeth and some lamellae 31 (maximally six). The maxillar head can be extruded through the mouth during feeding. The ungular teeth may aid the mandibular apical teeth in breaking down food particles, but the function of lamellae is far from clear. Goto (1972) suggested that the long and rake-like structure of Lamella 1 (median appendage) can function as a delicate rasping or collecting organ functioning rather like a bamboo garden rake. Lack of this structure in both S. henshawi and 0. hexfasciata indicates that these two species may have other adaptations for food collection when compared to I. notabilis. Labrum The labral shape of these three species is very similar, more or less quadrilateral with lateral margins basally converging to the apices (Figure 7). There is a deep furrow separating labrum from the fronto—clypeal region. Just behind the deep furrow there is a row of setae projecting forward and downward over the base, with four setae in I. notabilis and 0. haxfasciata and six in S. henshawi. All labra bear three rows of setae. The proximal row has five and the distal row has four setae in all three species. There are five setae in the middle row in I. notabilis and 0. hng’asciata, and four in S. henshawi. The setae in each row are very similar to each other, except for S. henshawi, where outer two setae of the proximal row are stronger than the three inner setae, and the outer two setae of the distal row arise from elongate rather than circular sockets. None of the labral setae seem to be modified as sensory receptors. The setae barely reach the distal labral margins. Beginning in the middle of the labrum three grooves (parallel to the main body axis) run towards the free margin in S. henshawi. These grooves separate four distal setae. There are four papillae on the distal margin of the 32 Figure 7. Labrum of three collembolan species. A. Isotoma notabilis B. Sminthurinus henshawi C. Orchesella hayrasciata I = proximal row. 11 = middle row. 111 = distal row. p = labral papilla. g = groove. 34 labra, and in 0. hexfasciata each papilla bears a very tiny seta. The estimated labral width of I. notabilis ranges from 31 to 55 um with 90% shorter than 50 um; 37 to 90 um with more than 94% between 50 to 90 um in S. henshawi; and 71 to 171 um with more than 94% wider than 90 um in 0. herfasciata. There is very little overlap in labral width among the three species (Figure 8). Again, it is difficult to relate the minor structural differences to the species’ food habits. The maximal mouth opening is limited by the labrum, and labral width may therefore determine the size range of particles each species can forage on. Number of Individuals with Gut Contents The percent of individuals of each species with gut contents ranged from a low of 64% to a high of 88%. Sminthurinus henshawi from both sites showed a slightly higher percentage than the other two species. Isotoma notabilis from Site I had the lowest percentage (Table 1). Three-way analysis of variance for unbalanced data was used to detect effects of species, site, week and their interactions for 1988; and of year (1988 vs. 1990), site, week and their interactions for 0. hexfasciata. There were no significant site effects with respect to the proportion of individuals with gut contents in 1988 (Table 2). Since species x week interaction was significant (Table 2), separate analyses of variance were run for each species after combining data from both sites. All three species showed significant seasonal variations (Table 3). The lowest proportion of I. notabilis with gut contents (Figure 9A) occurred in the middle of July (Week 11), Tukey’s test confirmed that it was significantly lower (P < .05) than in Week 1 (early 35 4O 1 l l ----- 0 hex/6504912? 30 _ ------------- 6? hens/)am' / notab/I/S‘ 1% 9 20 a ~ 0) (L 10 - ”\\ \ — ‘ [J \ ,’ \\ .2 J ‘t \ :- ’ \\ a '. I '7’] \A /\/I:' \x O “U‘ ”““‘i “““““ i ““““ O 100 150 200 Labrum Width {0) Figure 8. Labrum width frequency distribution of three collembolan species. 36 Table 1. Average percent of individuals with gut contents. Total Individuals Percent Species Site/ Year Individuals with Gut with Gut Examined Contents Contents I. notabilis* I/ 88 361 230 64 II / 88 1240 875 71 S. henshawi I / 88 381 303 80 II/ 88 749 656 88 0. hexfasciata I / 88 259 176 68 II/ 88 321 222 69 I / 90 294 226 77 II / 90 564 384 68 * Adults only. 37 Table 2. Analysis of variance of the percent of individuals with gut contents in 1988. Tested are effects of Species (I. notabilis, S. henshawi and 0. hafasciata), Site (I and H), Week and their interactions. EFFECT SS DF MS F—ratio P Species 30. 325 2 15. 162 12.690 .000 Site 1. 104 1 1.104 .924 .337 Week 8.373 6 1.396 1.168 .323 Species x Site .383 2 .192 .160 .852 Species x Week 131.151 12 10.929 9.148 .000 Site x Week 2.991 6 .498 .417 .868 Species x Site x Week 19.709 12 1.642 1.375 .175 Error 463.576 388 1.195 38 Table 3. Analysis of variance of the percent of individuals with gut contents in 1988. Data from Site I and H are combined. a. Isotoma notabilis EFFECT SS DF MS F -ratio P Week 23.947 6 3.991 4.280 .001 Error 79.272 85 .933 b. Sminthurinus henshawi EFFECT SS DF MS F -ratio P Week 143.823 7 20.546 19.496 .000 Error 219.201 208 1.054 c. Orchesella hexfasciata EFFECT SS DF MS F -ratio P Week 50.493 7 7.213 4.619 .000 Error 262.359 168 1.562 39 120 - 120 ~ A B 100 i 100 - 80 - go _ ‘5 7?. 8 so - 9 60 ~ 0) a> CL CL 40 r 40 - 20 - 20 .. O - — O - - 1 3 11 151719 23 1 5 9 1113 1719 24 Week Week 120 - 120 - C D 100 - 100 ~ 80 - 8O _ ‘c‘ E g 60 - 8 60 - cf 8 40 a 40 - 20 - 20 - O a _. O _ _ 1 5 9 11 13 17 19 24 1 5 9 13 17 19 24 Week Week Figure 9. Percent (means :1; standard errors) of individuals with gut contents for the three species from each site. A: I. notabilis; B: S. henshawi; C: 0. hexfasciata 1988 and D: 0.hexfasciata 1990. Filled bar: Site 1; Empty bar: Site H. 40 Table 4. Analysis of variance of the percent of 0. hexfasciata with gut contents. Tested are Year (1988 and 1990), Site (I and 11), Week and their interactions. EFFECT SS DF MS F-ratio P Year 5.573 1 5.573 4.028 .046 Site .211 l .211 .152 .697 Week 29.949 6 4.992 3.608 .002 Year x Site 1.203 1 1.203 .869 .352 Year x Week 16.790 6 2.798 2.023 .062 Site x Week 25.439 6 4.240 3.064 .006 Year x Site x Week 11.384 6 1.897 1.371 .225 Error 474.5 83 343 1.384 41 May) and 3 (late May). In early May and June (Week 1 and 5) S. henshawi had the lowest proportion of feeding individuals, but from early July (Week 9) until mid-October (Week 24), proportions were higher and constant (Figure 9B). For 0. hefiasciata, the seasonal change was not very clear (Figure 9C). Tukey’s test revealed that estimates from Week 9 were significantly lower than those from Weeks 5, 11, 17 and 19. The percent of 0. hexfasciata with gut contents was tested for 1988 and 1990 (Table 4), as well as for 1988 and 1990 separately (Table 5). In 1988 there were no significant differences between Site I and H, but in 1990 seasonal changes differed significantly between sites (Table 5). In Site I the highest proportions occurred in Week 1 and 13, but in Site 11 they occurred in Week 5 and 17 (Figure 9D). Table 5. Analysis of variance of the percent of 0.hexfasciata with gut contents in 1990. Tested are Site, Week and their interaction. EFFECT SS DF MS F -ratio P Site 1. 397 1 1.397 1.066 . 303 Week 15.226 6 2.538 1.937 .077 Site x Week 26.388 6 4.398 3.355 .004 Error 264. 825 202 1.31 1 42 When analyses were performed for each site separately, results revealed that there were no significant effects of year, week and year x week interaction in Site I (Table 6a), but in Site H seasonal proportion changes were significantly different between 1988 and 1990 (Table 6b). Food Particle Size Distribution Average frequency of food particle sizes in these three species is shown in Figure 10. The means of the large particles differed between species. In I. notabilis, most large particles were in the range of 44 to 110 um2 with very few larger than 110 umz. The majority of large particles found in S. henshawi’s guts ranged from 44 to 180 umz, and the largest was approximately 820 um2. Orchesella hexfasciata contained relatively evenly distributed large particles from 44 to 360 um2 with some extremes reaching 3600 um2. It is clear from Figure 10 that I. notabilis ate a high percentage of small and medium (both more than 40%) and a very low percentage of large particles (less than 15%). The three size categories’ distribution was relatively even for S. henshawi, with medium size particles only slightly more frequent (40%) than the other two. Orchesella hexfasciata contained a very low percentage of small particles (less than 20%) and a very high percentage of large particles (more than 45%). Seasonal variations in particle size distribution in each site are illustrated in Figure 11 for I. notabilis, Figure 12 for S. henshawi, Figure 13 for 0. hexfasciata (1988) and Figure 14 for 0. haxfasciata (1990). Isotoma notabilis always contained a very low percentage of large particles (less than 15% except Week 1), and small and 43 Table 6. Analysis of variance of the percent of 0.hexfasciata with gut contents in each site. Tested are Year (1988 and 1990), Week and their interaction. a. Site I EFFECT SS DF MS F-ratio P Year 5.149 1 5.149 3.434 .066 Week 13.148 6 2.191 1.461 .196 Year x Week 10.812 6 1.802 1.202 .309 Error 206.929 138 1.500 b. Site 11 EFFECT SS DF MS F-ratio P Year .952 1 .952 .729 .394 Week 47.691 6 7.949 6.088 .000 Year x Week 19.482 6 3.247 2.487 .024 Error 267.653 205 1.307 44 50 - S S S 40 — § S S ‘E 8 30 - ci’ Cl small I medium 20 § large S \ S 10 \ / 00/30/73 .5? hens/7a m’ 0 harem/ale Figure 10. Average food particle size (means i standard errors) distribution of three collembolan species. Small: < 6 um2. Medium: 6 to 20 umz. Large: > 20 umz. 45 70 - Site I 60 - 5O - E 40 - i3 at 30 - 20 e .. large 10 I medium 0 __ . Cl small 70 - . Site ll 60 - 5O - E 40 - 0 8 83 30 e 20 - .. large 10 I medium 0 __ . - 1:] small 1 3 ll 15 17 19 23 Week Figure 11. Seasonal food particle size (means 1: standard errors) distribution in I. notabilis. 46 7O - Site I 60 - 50 - E 40 - G) 9 i’ 30 - 20 - 10 _ large I medium 0 -- _ El small 70 - . Site ll 60 - 50 - ‘5 4O - (D 8 8: 30 - 20 - 10 _ large I medium 0 ._ _i _ El small 1 5 9 11 13 17 19 24 Week Figure 12. Seasonal food particle size (means i standard errors) distribution in S. henshawi. 47 70 - Site I 60 - I 50 - S S 4-0 S \ c 40 - S Q 3 l i (1’ 30 - § N \ 20 - E s - 5 large 10 S - g I medium 0 _.- _ _ S _ _ * _ Cl small 70 - . Ste H 60 - t 50 - E “a 40 - E ‘D S e S ci’ 30 - s 20 - 10 _ large I medium 0 _.. __ _ _ _ 1:] small 1 5 9 11 13 17 19 24 Week Figure 13. Seasonal food particle size (means i standard errors) distribution in 0. hexfasciata, 1988. 48 70 - 60 Site I 50 - E 40 - Q) 8 83 30 - 20 - 10 _ large I medium 0 J- _ [:1 small 70 - . Ste H 60 - /////////////////)—-' Percent 50 - 40 - 30 - 20 - 10 _ large I medium 0 ._ _i _ Cl small 24 1 5 9 13 17 19 Week Figure 14. Seasonal food particle size (means i standard errors) distribution in 0. Wasciata, 1990. 49 medium sizes usually ranged from 40-50% each. In most cases, each of the three size categories constituted 20—45 % of the total in S. henshawi, with medium size particles most prominent. For 0. hexfasciata, the percentage of large particles was always higher than 40%, sometimes near 60%, and small particles were infrequent (usually less than 20%). Multivariate analysis of variance was used to test differences between the species in 1988 (Table 7) and differences between 1988 and 1990 for 0. heifasciata (Table 8). Results showed highly significant differences among species and among weeks. Separate analyses of variance for each species and year (Table 9) revealed that there were no average differences between sites for either I. notabilis or S. henshawi; however, ‘ significant interactions suggested that particle size distribution underwent different seasonal changes in the two sites. In 0. hexfasciata, particle size distribution did not differ between years (Table 8), and there were no site effects in either 1988 or 1990 (Table 90 and d). Food Components a. Average dietary components of each species There were general similarities between species: fungal hyphae and spores, plant materials and colloidal materials made up the majority of gut contents (Table 10). On average, more than 95% of gut contents of I. notabilis and S. henshawi, and more than 80% of these of 0. hexfasciata consisted of these four food items. 50 Table 7. Multivariate analysis of variance of food particle size distribution in 1988. Tested are Species, Site, Week and their interactions. EFFECT Wilk’s Lambda F-stat. DF P Species .277 162.175 6, 1082 .000 Site .998 .353 3, 541 .787 Week .895 3.393 18, 1530 .000 Species x Site .995 .450 6, 1082 .846 Species x Week .782 3.856 36, 1599 .000 Site x Week .977 .710 18, 1530 .803 Species x Site x Week .876 2.032 36, 1599 .000 Table 8. Multivariate analysis of variance of food particle size distribution in 0. haxfasciata. Tested are Year, Site, Week and their interactions. EFFECT Wilk’s lambda F-stat. DF P Year .991 1.085 3, 359 .356 Site .995 .585 3, 359 .625 Week .737 6.435 18, 1015 .000 Year x Site .996 .497 3, 359 .685 Year x Week .924 1.603 18, 1015 .053 Site x Week .915 1.808 18, 1015 .020 Year x Site x Week .955 .928 18, 1015 .544 51 Table 9. Multivariate analysis of variance of food particle size distribution in each species and year. Tested are Site, Week and their interaction. a. Isotoma notabilis EFFECT Wilk’s Lambda F-stat. DF P Site .979 1.318 3, 188 .270 Week .640 5.051 18, 532 .000 Site x Week .845 1.810 18, 532 .022 b. Sminthurinus henshawi EFFECT Wilk’s Lambda F-stat. DF P Site .989 .837 3, 215 .475 Week .647 4.817 21, 617 .000 Site x Week .839 1.852 21, 617 .012 c. Orchesella hexfasciata 1988 EFFECT Wilk’s Lambda F—stat. DF P Site .999 .050 3, 184 .985 Week .756 2.582 21, 528 .000 Site x Week .877 1.175 21, 528 .267 d. Orchesella hexfasciata 1990 EFFECT Wilk’s Lambda F-stat. DF P Site .988 .852 3, 201 .467 Week .678 4.652 18, 568 .000 Site x Week .880 1.460 18, 568 .099 52 .H 55 32 H... 3533 25c ”AZ jam .2 .11; film Sum Hum meow mu; Tire we o3: Hue H... .11; He: New See mesa mama Ham a: 8: Tire .2 *ul «HS. Sam was meow «Mom 7.: me at: an... an .11; Nam TIL. was me: mama lab 2: 3: 8398330 in; .2 .31: Nuom .31....71: “HSNHSHH: 0: SE *9... m2 .3: _HwN m2 7:; Ta: me? Ill: w: a: Mercedes.” .13 “Hz 1.... Tum... dz .2 He: Has mama we at: Tum .2 *u... News m2 m2 _He Sea «a; 8 mm: ceased 6:5 Sam 5&2 34.8 .85 doe a2: 8% an? .22 Sign ”28% .«o H.585: 38 H .Gofiwao maoemoomm .oZV deflonsszoo 8:: we 35w 2: 5 35:09:00 E06 E Echo ENE—Sm H 2.8.5 Eugen .2 2an 53 There were also striking differences between species: in each the major food items appeared in different proportions. Colloidal materials and fungal hyphae were the dominant foods in the gut of I. notabilis. In S. henshawi and 0. hexfasciata, fungal spores were most frequent. Pollen was another important food item for 0. hexfasciata, especially in 1988. Most plant materials found were brown, unstructured and humus-like substances. Only a few 0. haflasciata contained litter remains of higher plants. Colloidal materials were mostly of unknown origin. They could have been partially digested food, or may have been in a liquid state before ingestion. Isotoma notabilis had the most restricted or specialized diet. It did not feed on pollen, animal remains or bacteria. Sminthurinus henshawi could feed on pollen and animal remains, but animal remains were found in only two specimens. Orchesella hexfasciata had the most varied diet. Due to method limitations, small amounts of bacteria could not be detected. Only two specimens of 0. hexfasciata were found with large quantities of bacteria in their guts; both came from Site I, one in 1988 and the other in 1990. b. Interspecific comparisons There were highly significant interspecific differences with respect to all food categories except bacteria (Table 11). Site effects (across all species) were never significant. Seasonality (week effect) was strong for all major food items except for colloidal materials. Significant species and week effects also resulted in significant species x week interaction. It was difficult to interpret the other interactions in a 54 ox. 3a. was. Re. 8m. 4%. E. 4mm. wmm. 383 x as x ”28% as. as. 3. 2w. :m. moo. me. So. 9e. #83 on am So. 3%. So. So. So. So. So. So. So. 983 x 83on So. as. as. so. me. E. 48. a. go. as x “28% So. as o8. me. So. So. So. o8. o8. gets me. we. mew. mew. E. «8. 09.. as. me. ea 80. war. 25. So. So. So. So. So. So. 328% $225 museum weave: a: .5 use... 5:8 a: .5 seem “seam Bmmmm .mcouofiBE :05 98 x83 Jam .8603 new @888 .32 E 2:0: e08 Co 8533 Co $3.28 we 829d .2 03m... 55 biologically meaningful way. Most of the pollen encountered in 0. hexfasciata was pine pollen, whereas more than 2/3 of the pollen found in S. henshawi was non—pine pollen (Table 12). Statistical analysis confirmed that these two species consumed pine and non-pine pollen in different proportions, and that there was no site difference for either species. 0. Potential year-to-year variation Analysis of data for 0. hexfasciata revealed that fungal hyphae, colloidal materials and animal remains were very constant food items in 1988 as well as 1990 Table 12. Relative percentage (based on the total amount of pollen) of pine and non—pine pollen found in the guts of S. henshawi and 0. hexfasciata. Species Site/ Year Pine pollen Non-pine pollen S. henshawi I/ 88 32.2 67.8 11/ 88 26.9 73.1 0. hexfasciata I/ 88 90.5 9.5 II/ 88 94.5 5.5 I / 90 91.6 8.4 II / 90 93.2 6.8 56 (Table 13). However, specimens from 1988 contained significantly more pollen and less plant material than those from 1990 (Tables 10 and 13). There was no difference in the proportion of pine and non-pine pollen between 1988 and 1990 (Table 12). d. Seasonal changes in dietary components Results of analysis of variance of food items for each individual species are shown in Table 14. In I. notabilis, most food items showed very pronounced seasonal variation, only plant materials being marginally significant (Table 14a and Figure 15). Compared to the other two species, however, seasonal variations in major food components were relatively moderate in I. notabilis (Figures 15, 16, 17 and 18). Seasonal (week) effects were also significant for most food items in the guts of S. henshawi (Table 14b). Fungal spore content increased sharply and plant materials decreased in the last week (Week 24) of the season in both sites (Figure 16). Pollen only occurred in spring and early summer. In 0. hexfasciata, proportions of animal remains as well as minerals showed significant seasonal changes in 1988 but were non-significant in 1990 (Table 14c and d). On average, animal remains only accounted for 4% of gut contents; in some weeks, however, (e.g., early July in Site 1, Week 9 of 1988) they reached 15% of total gut contents (Figure 17). Mite and collembolan exoskeletons (mainly 0. hexfasciata’s own cuticles), protozoa and nematodes were found among these animal remains. Fungal spore content increased sharply and plant materials decreased in the last week of both years (Figures 17 and 18). Orchesella hexfasciata fed on pollen only in spring and early summer but not at other times. moo. owm. mam. moo. mom. bmo. who. mom. mmo. x8? x 85 x 50> owe. owm. :uo. ooo. mew. 3m. mom. moo. mow. “48>? x 8% mo“ owm. oom SN. mmo. ooo. me. ooo. moo. #83 x 58> omo mom. owo woo. vow. bmo. mew. Ev. mom. 8% x 53% ooo. owm. boo. ooo. ooo. ooo. ooo. ooo. ooo. #83 one. mom. woo. “mo. mmm. oww. wmh. woo. woo. cam emw mom. H2 Ohm. woo. ooo. ooo. omo. 3w. Eu? :BoEED «€8on macaw/H .32 EU REE< cozom .32 .NE 88mm macaw Hommmm .EBSSBE .55 28 x83 .86 new.» 8e c989 .SEQSKSQ .O E was: 38 E 85E; E maiden E 833$ .2 2an 222 moo. 28>? x 8% .32 .EU REEF. So. 33 4mm. 2% saga: 52mm 58 .2333 22.232326 .9 222 222 222 a: so as? RN. 0283 x 2% 48. 328% e8. 8% SEE 52mm £328 33:5 .m .Eomoa use: H AZ doEfiBE 22: 28 28>? .25 oh" 833. .8609. 8:232: E 25: Eco E 8253 E 29:28 E Soweto .2 29mm. 59 Em. at 2%. So. an. wmo. 5. am. one. #83 x 36 So. 5% EV. 8o. 1;. So. So. So. Re. #83 can. 8m. am. 95 E. 2:. o8. So. Nam 35 5852: 838m @052 .32 .28 355. :23 .32 a3 88% 022m Emmmm 82 32983 33855 a «mu. am. so. ms. 8N. mom. Rm. So. so. M83 x 9% ooo. Rm. m8. Re. So. So. 25. Bo. Bo. V5.; m3. 8v. be. on. 0%. 8m. 8m. 8m. NE. 95 535:5 «E823 mfiuocuz .32 .30 Bang =o=om .32 .wE macaw 039mm Hum—"Em £2 assahfi §m§§c .o $35283 E 2an E COIIOIdaJ materlal E unknown El mlnerals s plant materlals E spores hyphae l wuwuwwwwmw”"ququu»...uuuunuwumwunuwuummwm § wwwwwwm \\\\\\\\\\\\\\ 1 L v............. hone???nonononowowouono ‘ .......l..l....i...... ouououououou.u.»ou.nou.uou.uou.uouon m .ww "WWW”«WWWw«WWWuwwwwwwmuwwwwuwwwm§mwmwmm.\\\\\\\\\\\\\\\\\\\\ . 0000000000004OOOOOCCQCOOOOOCCOOO€ 366‘ ooo.ooo.oooooooooooooooooooooooo \\ .3 \\\\\\\\\\\\\\\ .0‘O©OOOOOEO©O©990wOOOOOObOOOOoowOO.0‘00.0000OOO‘O‘O‘O‘O‘O‘OOOOO‘O‘O. ‘ s "ononou \ 1 l l I ll [[11 . pwmwmwwwwwmumwm»"wwwmuwwmwm§wwwwme\\\\\\\\\\\\\\\\\\ m nmuwwwwwwmuwwwwwmmuwwwwwwmV\\.w§\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ _ _ 4 2 1 OO 80 60 asp—ma «nun V\\\\\. wwwme\\\\\\\\\\\\\\\\\\\\\\\\\\y 15 17 19 23 11 O O O E colloldal materlal E unknown El mlnerals s plant materlals B E spores hyphae I. mmmmwwwmwwmmmmfiuwwwmwmwmmm”.V\\\\\.mwww\\\\\\\\\\\\\\\\\\\\\\\\ I1 1 «a .qo.%o.oq%%%oco«flodcoqocoqfiocoqoco...04..N ............ \\\\\\\\\\\\.\\\\\\\\\\\\\ «nu homewovwowowowowowowowowowowowowowowowowo.040. ‘ \\ "ununuuuuuw \ \ \ , W u”unuuuuwumuuuumuumm»NMNMNNNMS wwwwwwwww\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ .a\\\\\\\\\\\\\\\\\\\\\\\\\\\ . “n” 94.{0«.¢.¢0o.¢0¢0¢.n040¢0«0....a0q0c01Oco.0¢.¢.«Oa.¢04. V n94 "u" hone».nonououononononenonououononononououononouonoNou \\\\\ .u ’0 O . I r I. V . .l... . .l. .4. . . . . O . .‘E...‘ . . . . . . . . . ‘ .cOcOQO ‘ 000000009o9009000900000.ooooo.\\\uo._ \\\\\\\\\\\\\ a k3....vofivvvv..vv¢..&.v£.£&.vvvvoo...\ :x...\\\ \ \\ wwwwwmmwwwwmwwwmww«mwmmmwm§Wwww \\\\\\\\\\\\\\\\\\\ m._ WWmMummuwwwmuwwwwwmmwwwwmmm g .\\\\\\\\\\\\\\\\\\\\\\\\ 15 17 19 28 11 Week Site 11. B : Site I. A is 1'] Figure 15. Seasonal food components of I. notab pollen plant materlals hyphae pollen plant materlals hyphae Site 11. B 61 a colloldal materlal E oolloldal materlal Cl anlmal remalns El anlmal remains E unknown I] mlnerals SS s m spores E unknown El mlnerals E spores § RV B — _ _ _ u — WOOOOOOOOOOOOOOOOQOOOOOOA \\\ 0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 \\ 0.0.0.0.0»0.0.0.. \ - 24 .‘O........... .O............O.................‘............... 0.0..» v .w.no?»0w040m0w0m0n0m0w0u0u0w0n.m0u0u0u0u0m0w0u0m0u0u0m0w0m0u0u0u0m0u0w0nom0n0w0w0u0n0u0n0w0u0u0n0w0 9b.».vorO>ObOvasr00¢>O>O>Ou \ 3&0»0w0wou0u0u0u0w0m0w0w0u0n0w0m0u0m0u0n50?».u0u0now0n0n0u0u0w0w0u0m0m0n0wow0u0m0u0u0m0uow. H U .. I L m .O.O{OCO{OCO‘Q‘O‘OCOQO‘C‘O‘.‘O‘O€ H0H0M0M0M0H0H0H0N0N0N0H0H0M0H0H0N0H0H0H0M0H0M0N0H0M0H0“ ‘\\\\\\\\\\\ — v.09...GO{O‘OGO‘O‘OCOQOQO‘OfiO‘OCQ §\\\\\\\\\ "0HOHONOMOHOHOMOHOHOHOH0u0MOMOMONOMOHOMOHONOMOHOMOMOHOH. § _.....wowfiowowovwowow$0?»$0»..o. ..K.»as.3.3.3.an???Shaman.” homomowowogowow.wowowomowowoa\ \.a.”.uaa.".3.“.3.wwwuanaaaaa.".55\ ‘9‘... 0000000000000 0‘ l ‘0‘ ‘01.. 1061.00.00 V a. ...l........‘0.... u 00000O01010401010...090.061.90.010...0.000.000.5100‘ O. O ‘ 000000000000 0000000000000000000. ...v .0 0000 O. .0000000000000000000000000000000.00000 O. OOOONOOONO. \ \ n0H0H0H0H0H0H0M0H0N0H0N00 0 "00000000000u000000000n0u0H0H0H0H0M0u. \ ... “NOHONONOOOOOWO.OOOOOOONOO O. ‘ k .M0H0H0H0H0M0H0H0N0u0n0 00000000000000000000000000000000000000000M0H0H0n0n0” \ . 0....00....O..0.0000 .0 17 19 ..... ... .................. ‘\\\\\\\\\\\ \\\\\\\ ......... . weevvvvva \ \. tx.“aua.u.x§$§.uexn \ ‘ “‘ “‘ .‘...... ...........4 ...".”."$... $3"... V\\\\ 0,0.0.0.0.0.0 . . . . l 0100000000000 00.0 00 . D s ‘x\ 91118 W OE .0000 000000000000000000000. 00000000000000000000000000000000 .‘..‘........‘...“4 n.3.x.x.".x.u.u.u.u.u.u.u.u.n.u. V\\\\\\\\\\\\\ 00.).bODObOD000000000000.00.OIODODODOD. “ I...‘..‘....... _“ § .ua.”.u.u.".“.".".“.u.u.u.u.n.. . . a t V0.0.0.05.0.0.0.0.0.0.0.0.0.0.0.0.0. . - l I I l “_1“‘.““‘ “‘ q0¢0¢000c0100010c0c0004u 0000000000... 00000000000 69 .000000000000\\\\\\\\\\ 0000000000000 ..~.......w.......w... ..... v".n.n.n.».u.u.n.n.u.n._ 5 .vvvvvvvxa.”.u.“ l I l I ‘.‘.‘.‘.‘.‘.{.‘.‘.‘.‘.{.fi \\ ‘\ V0.“.1.1.1.¢.1.¢.1...1.1.<.N.M.M.“OC.4.¢O¢ONO. ‘ w».».».».».».».» R.» u\\\\\\\\s § «Hang»”mmuNnun».n.n.u.»"»~.u \ 1. 9004.:00.4.1O4O{O0>0v0>0>0>0p0v0> .00. L ...0.0.0.0.0.0.0.0.0.0...05.03.0.0.90.90.0.0.0.0.0.0.05.90.90.39..05....3.990.033.9059 2 l l l l v. ““...“.““““A 000.00000000000000000000.. . ....““.‘ u 00000000010000.0000. 0 ‘ 0000000000000000000000000.‘ . 000000000 0000 0 00.‘ an wWu»«wuwnwuwnwnwuwuwuwuwuw.wuwwuwg I g wumuunuuua»umwwuu»"muwwwwwwwwnmmu». \\ 0 0 ..0....0u~uwuwu~um_ g manuwwwumnwumwmummwwnmwwu \\ 17 19 17 19 .0 0 0 0 0 0 0 0 0‘0. ‘ . 0000000000000000.0.. ‘ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 0.000 0.00 00000M0000000000. V ...».u.».».».».n.».u.. \\\\\\\\\\\\\\\\\k . . .u.".a.n.u.”.u._ \\\\\\ ....u.».u.n.u.u.».».u.».na.” \\\\\\\\\\\\\\\ na”"nan.”«uam.».u."nae”. \\\\\\ I. 0|. I. I. M0». “““““‘ .‘ .‘......................“‘........a . _...................... I “\\\\\\\\\\ .3"a3a.aa.x.a.aa.n.n.n.u.”.".".”.§fig.".3 V\\\\\ ..... .p0p0p0p0p0p0. .0.0.0.. ‘1 L 00.0.0.0l010.0.00.0.0.0000.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.00.. ..‘......"......‘. ‘ tx.a.a.".".”.".“.a.".u.u.u.3x.“ \\\\\\\\\ 0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0,0 \ \ 0 0.0 0 0 0 0 0 0 ‘ 0.0.0.010.0.0.0.0.0.0.06.0.0.0.010.00.010.06.0.0.0.0668 . . § 00000000000000000 I \\\\\\\\\\ 00000u0000000000000u00000H0u0u0000000u0N00000000000n00000000. ‘\ A . \ I.........‘. 0.0; ‘ 00000 0000000000000 00000 0 0 0000000 0 00000000000 000000000 . au00001.0014«1004.014141411 11111 . “.33... _” V\\\\\\\\\aex.a.a."a.a3".u.u.n.n.n.u.n.u.n.n.uaka.“§\ .......».,.. , ‘ .000000000000000000000000000000000000000000000000000000000000000 91113 W 91113 W.unawawwwwml§V\\\\\\\\\\\\\\\»mmma¥\\\\\ umawwamlm kmmmmmmmma.\\\\\ m w_“Manama“"mamIs\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\t§m‘ 5 m” gamma a_amMM“wwwwwwwwmama §§§WMWMWMWMWMWMWMWMWMMV 1 .wwwwwww§\\\\\\\\\\\\\\\\\\\\§§amammama? 1 _ _ 7 P _ _ _ _ mm m m m o m o o m o o Eooaoa #588 Site 11. B 'te I l S A 1988 W Figure 17. Seasonal food components of 0. hexfasciata, 63 % m WWW E E m .m SS S E colloldal materlal El anlmal remains pollen mm mm ED plant materlals E unknown bacterla D mlnerals E unknown & bacterla El mlnerals E spores hyphae A B _ _ — _ _ _ _ — ‘ “{““““{‘ '0““““‘00""‘.“““““‘0.0.““ “““‘ “““{““‘-‘ .0..“““..“.‘.‘{‘ ““ ““‘ 0 “ “00 . .00 00000000 .‘ 00000000000000000000000000000000000000000000‘ 4 . 0 0 0 .‘ 00000000000000000000040000010000000001000000000000004. 4 v . ... . . O O . O . O Q \ 00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000A \ . V.’ 0.. ............. ..... \ 0000.00000000000000000000000000000...0 00000000000000000000000000000000000"000000000000000” \ . V0>0v0>0b0v0>0>0p0p0>0b0n . 30.0.0.3..0.0.0...0...0.0.0.0.0.0.0.0.0.0.0.0.0.0.0...0.0.0.0.0.0.0.90.05.03.05»?! 2 . r0b0>0v0> >0b0b0>0r0>0b00 s. .0.05.0.0.0.990.003.9050?.0.0.0.0.0.0.0.0.0.0.0.0.0.0.9003310.30.890.90... \ 2 l l I l l l 0 0 0 0 0 0 0 0. 00000000000000000000 0000000000000 00 . ”Keven.” I g a?”.a»mamwwamaw V "0”. . m 0 . . .l. . . . . . . C ‘ 00.... 0 . .0...l... 0'..l...1.,..r..d‘ ‘ a. n.».».u.n.».».».».n.u.l \\\\\\\\\\\\\\x.".3""a”.nawauua.”“"3".“\\\\\\\\\\ . 0000000000000‘000 \ 0000000000000000000 00000000000000000\\\\\\\\\\\\\\\\\\\000000000000000000 V v00.00.00.0w00.0w00.090.00.0w00.00.00,0w00.0 \ nunnumun”anunannouuunuuuuuuuumuuuuunu \ 17 19 17 19 .5 00000000000000000 l l l l l l "mumuwwmmnwwww I g "wwwwwwwwmmawwmmuwwwwwwm § mm m _ wwwmumeuwwmeuwmm I a g mMMnwwwmmMWmMMamawwwwmmuwwwwmM .\\\ , mm 1 W l l I I I v4.1010«0.0.0.0¢0.010.0«0«0.040¢040. V V .0000000000000000000 V\ 04..0cO.0<.101.1.104..0¢.a.¢.¢. \ 0000000000000000000000000000040000. ‘ non0nouou.n0n.uonon.n0n.n0n.n0n.u. I \ ”gamma”? \ 9 _v».n.n.n.n.n0u0n.n0n.”0”.» no». I m .muwwwwwwwwwwwfimmm \\\\\\\ 9 wwwww I § § mmmmmmmmmmmmmmmmmmmmm \\\\\ . .new.Is\\\\\\\\\\\\\\\\\\\\\\\s§mamwemmmmmmmwmManamaw 1 m.m.wm.wwmmls\\\\\\\\\t . \\\\\\\\ mm m m m o ucoobm nu 20.0.0.06.0.0.0.0«0.0.0.0.0.0.0.. \\\\\\\ E .w.~.w.«.n.u.m.«.m.~.. X V".~_3.n.».».en.n.vu n.x.x.u.u|\ \a\ sw.n.”.n.".u.”.n.u.u.a\ — Site 11. O O B 'te I 1 S A W Figure 18. Seasonal food components of 0. hexfasciata, 1990. 64 Discussion There is always some proportion of a collembolan population lacking gut contents at any one time. Joosse and Testerink (1977) reported that under both field and laboratory conditions, only 50-60% of Orchesella cincta (Linné) had food in their guts, and that low temperatures and long periods of drought tended to decrease the number of individuals with gut contents. Long periods of drought depress the growth of soil and litter microorganisms, which become unavailable or inaccessible as food for Collembola. The year 1988 was very dry, and rainfall was distributed unevenly over the season. From May to July precipitation was well below the long-term average; significant rains did not begin until the end of the first week of August (corresponding to week 13) (Snider and Snider, 1989). Precipitation in 1990 was relatively evenly distributed, and rains were ample during May and June (Snider and Snider, 1991). The prolonged drought in 1988 may have reduced food availability for some animals: I. notabilis had a relatively high percentage of colloidal material in its gut, and the low proportion of individuals containing food in Week 11 (Figure 9A) can thereby be explained. It seems that S. henshawz' was not significantly affected by drought with respect to the proportion of individuals actively feeding. In Week 1 and 5 a low proportion of the population had gut contents, but in Week 9, a very dry week, the proportion was increased (Figure 9B). For 0. hexfasciata the proportion of individuals containing food in Week 9 of 1988 (Figure 9C) was statistically lower than in Week 9 of 1990 (Figure 9D), possibly also due to the long period of drought in 1988. In Week 11 of 1988 the proportion of 0. 65 herfasciata with gut contents increased, concurrent with increased litter moisture. The discrepancy in feeding activity between I. notabilis and 0. harfasciata in Week 11 could be due to the different sampling methods and/or feeding habits of the two species. The long periods of drought have different effects on different collembolan species. Joosse and Testerink (1977) also reported that Collembola extracted in Tullgren funnels had an extremely low proportion of animals with food in their guts. In the present study, I. notabilis extracted in Tullgren funnels yielded a reasonable number (average 69%) of animals containing food compared to previous work (SO-80% with gut contents). The majority of Collembola in a given litter sample were obtained during the first day of extraction (data from a preliminary study). Extraction of all samples was carried out under the same conditions, so that loss of gut contents, if any, should remain relatively constant over different sampling dates. The Tullgren funnel extraction method used here appeared not to have strong effects on variations in gut contents of these animals. Pitfall traps (a method which relies on active movement of animals) usually yield a higher proportion of individuals with gut contents than other methods, because molting animals not only do not feed, but are also inactive. For individual species, the potential bias of high proportions with gut contents should remain constant over the year, and results can be validly used to indicate the feeding behavior of these animals through the seasons. There is ample evidence from aquatic insect studies that coexisting species consume different foods (Wallace and Merritt, 1980). Larvae of three trichopteran species utilized different sizes and types of food to partition their resources (Wallace, 66 1975). Mouthpart structure was found to be closely related to the feeding habits of Diptera larvae (Surtees, 1959; Harbach, 1977). In a study of soil oribatid mites, Kaneko (1988) concluded that the shape of the chelicerae was related to feeding habits of these mites. In this study, the three dominant litter-dwelling collembolan species also partitioned their food resource by utilizing different types and sizes of food. Food size partitioning was determined by mouthpart structure of each species: animals with large mandibles and labrum ingested a greater proportion of large particles, animals with medium-sized mandibles and labrum ate a larger proportion of medium-sized particles, and animals with the smallest mandibles and labrum ingested mostly small particles. However, these " chewing type" species may use their mouthparts to chew and tear food particles before ingestion, and the size of particles found in their guts could be skewed towards the smallest particle size fractions. Results of this study still show very strong evidence that the three coexisting species partition their resources through food size. The largest particles found in these collembolan species were animal cuticles and pollen grains. Both were found only in S. henshawi and 0. hexfasciata but not in I. notabilis. Lack of pollen in I. notabilis indicates that I. notabilis lives in a different microhabitat compared to the other species, and that pollen may be unavailable to it; or that I. notabilis does not prefer pollen; or that I. notabilis cannot handle pollen because of the small size of its mouthpart structures. The labral width in I. notabilis ranges from 31 to 55 um with 95% of the measurements smaller than 50 um. The maximal mouth opening for I. notabilis is thus less than 50 um, minus the spaces occupied by mandibles and maxillae. Pine pollen found in the field measured 80 x 40 um, and non-pine pollen 67 measured about 40 um in diameter. Thus, it is most likely that I. notabilis did not feed on pollen because of size restriction. Orchesella hexfasciata has the biggest mandibles and labrum, and its gut also contained the highest proportion of animal cuticles and pollen grains. Presence of different food types in the animal gut does not necessarily mean that the animal derives its nutrition from these items in proportion to their occurrence in the gut. But a close relationship between ingested and assimilated food would be expected from evolutionary processes (Cummins, 1973), since ingestion of large quantities of non- assimilable food would represent wasted energy and would reduce the competitive potential of a species. The method employed in this study of gut contents of Collembola has a well- known weakness: it tends to over-emphasize the relative importance of durable, non- digestible materials, and under-estimate the importance of quickly-digested materials. However, gut content analysis is one of the techniques most readily available and widely used for the study of food selection and partitioning. These three species also partitioned their resources by using different food items. Isotoma notabilis fed more heavily on fungal hyphae and colloidal material, S. henshawi ingested a greater proportion of fungal spores and 0. hexfasciata incorporated more diversified foods in its diet. The gut contents of I. notabilis in this study agree in general with a previous study (Poole, 1959), in which fungal hyphae were found to be a more important component than spores. Although both pine pollen and non-pine pollen were found in 0. haxfasciata and S. henshawi, each species preferred pollen of a different kind (Table 14). 68 Increase of spores and decrease of plant materials in the gut of 0. hexfasciata and S. henshawi in late fall could be attributed to increased availability of fungal spores and decreased palatability of leaf litter. Production of fungal spores follows a very strong annual cycle connected with the periodicity of external conditions (Ingold, 1971). It usually peaks in late summer and fall. Rastin et al (1990) found that fruiting-body formation by litter-decomposing fungi in a spruce forest floor began in July, biomass production remained low during the summer months, increased considerably in fall, and reached its maximum in October and November. Based on both field and laboratory work, Knight and Angel (1967) believed that fungal spores were preferred over all other food types by Tomocerus flavescens, and that the low spore content in field animals was due to lack of spores in the field. The field animals they analyzed were collected between mid-June and mid-August, and the present study also showed relatively low fungal spore content during this period. Newly fallen leaves are not a preferred food for most soil animals; they must be conditioned for some time before the animals will ingest them (Seastedt, 1984). It has been demonstrated that the longer leaves have been on the ground, the higher the rate at which they will be consumed by Collembola (Sadaka and Poinsot-Balaguer, 1987a and b; 1989; Sadaka et al, 1988). Schaller (1949) also indicated that Collembola would not feed on fresh leaves. In the study sites, peak abscission occurred during September and October (Snider and Snider, 1989, 1991). These freshly fallen leaves would not serve as a food source for surface-dwelling 0. hexfasciata and S. henshawi. There could be two possible reasons for the appearance of plant materials in the guts of S. henshawi and 0. harfasciata: 1) when the animals feed on fungi, they 69 accidentally ingest the plant materials connected with fungi; 2) plant materials serve as secondary food source when fungal spores become scarce, providing some nutrition. Fungal spores found in these three species consisted of a large number of species. Different fungal species have been related to different decomposition stages of leaf litter and to leaf litter species (Kendrick and Burges, 1962; Hudson, 1968; Widden and Parkinson, 1973). In food preference experiments, it has been demonstrated that Collembola selected some fungal species over others. Unfortunately, identification of fungal species found in collembolan guts was beyond the scope of this study. Pollen has been shown to be a highly valuable food for a large number of insects (Kevan and Kevan, 1970). It contains high amounts of protein, carbohydrates, lipids and various vitamins (Lunden, 1954; Lubliner-Mianoowska, 1956). Although Entomobrya comparata Folsom was observed on the pollen cones of pine trees during their flowering season, there was no evidence of 0. hexfasciata visiting pollen cones to feed on them. Pine pollen matures only in May and June and is wind-disseminated in great abundance (Barnes and Wagner, 1981). The availability of pine pollen is thus seasonal, as is its appearance in the collembolan diet from May to June of each year. Pollen must play an important role in the nutrition of these Collembola, especially of 0. hexfasciata during spring. Scott and Stojanovich (1963) found intact juniper pollen in the gut of Onychiurus pseudofimetarius Folsom in various stages of digestion. Both intact and broken pine pollen were also observed in the gut of 0. hexfasciata, intact pollen measuring about 80 x 40 um. It was clearly indicated that 0. herfasciata " swallows" these foods instead of chewing them before ingestion. 70 There were pine trees in the area surrounding both sites, but not in the sites themselves. Dissemination distance of pine pollen is affected by environmental conditions. Higher humidity during dissemination would limit the distance pollen can travel. The difference in percent of pollen in 0. hexfasciata guts between 1988 and 1990 could be related to the amount of rainfall during dissemination. There could have been more pollen distributed over both sites in the drought year of 1988 than in the relatively wet year of 1990. Pollen was thus likely to be more available to collembolans in 1988 than in 1990, and gut content analysis showed a high pollen content in 1988. The three collembolan species living in or on the forest floor probably occupied different microhabitats. It has been long realized that Collembola are morphologically adapted to a greater or lesser degree to either a subterranean or surface-dwelling mode of life (Poole, 1961). Species with pigment distributed in patterns; 8+8 large eyes; strong furcula; long antennae, legs and setae; complicated sensory setae on the tibiotarsae; and lack of postantennal organ are considered epigeic forms (including most entomobryids and sminthurids). Sminthurinus henshawi can be found in above-ground habitats (Snider, personal communication). On the other hand, I. notabilis represents the transitional form from hemiedaphic to euedaphic, being absent from above-ground habitats, and dwelling mainly in litter and humus layers (Bockemfihl, 1956; Poole, 1961). In the present study, I. notabilis frequented both leaf litter and A horizon, whereas the other two species, S. henshawi and 0. hexfasciata, were clearly surface-dwelling and did not frequent soil. In summary, these data suggest that size as well as type of food particles consumed are important in niche segregation of litter—dwelling Collembola. In two 71 similar but spatially separated forests, each collembolan species foraged on very similar foods, in terms of food type as well as particle size. Species-specific mouthpart morphology determined the particle size ingested. Gut contents of each species also reflected seasonal variations in available food type and the influence of environmental conditions. All together, these observations document mechanisms for reducing interspecific competition for food among species occupying the forest floor. CHAPTER II FOOD PREFERENCE AND EFFECTS OF FOOD TYPE ON THE LIFE HISTORY OF SONIE SOIL COLLEMBOLA 72 73 Introduction In a food preference experiment, Singh (1969) found that three collembolan species, Tomocerus longicomis, Onychiums armatus and Neanura muscorum chose different foods. They preferred fungal colonies and humus although they could feed on a variety of foods. Onychiurus armatus preferred fungi with smaller spores and finer hyphae while T. longicomis fed mainly on larger spores and hyphae. Neanura muscorum selected fine particles or minute spores in suspension. Both T. longicomis and 0. armatus also accepted bacteria as food but when a fungus was offered as alternative, it was preferred. All species could feed on humus and decaying wood but refused to accept them after they were sterilized. In a study utilizing Hypogastrura tullbergi (Schaeffer), and 43 fungi and actinomycetes, Mills and Sinha (1971) found that the fungi which were consumed most readily and permitted the most rapid reproduction were generally those with a low spore count and a mycelial mat that allowed free movement of Collembola. Since the authors conducted their study by introducing Collembola into agar slants with a thick mat of mycelium and spores grown over the surface, their results were somewhat biased. The animal neither fed nor reproduced on potato dextrose agar offered as a control. Shaw (1988), in a laboratory study, showed that Onychiurus armatus formed a perfect hierarchy of fungal feeding preference. Schultz (1991) observed that two isotomids, Folsomia candida and Proisotoma minuta (Tullberg) had the same preferences for ectomycorrizal fungi. There were some differences between multiple and pairwise 74 choices. It has been reported that Collembola not only had a preference for certain microbial foods, but also exhibited selectivity between different components and growth states of a microbial food (McMillan, 1976; Visser and Whittaker, 1977 ; Moore et al, 1987). In food preference experiment, McMillan (1976) found that Onychiurus armatus accepted all species of yeasts and fungi investigated, but preferred spores of one fungus, and hyphae of another fungus in multiple choice experiments in which sporulating and non-sporulating forms of two microbes were presented. Both microorganisms had similar morphology. The cause for selective grazing by 0. armatus was not obvious in the study. Onychiurus subtenuis Folsom preferred pigmented fungi over hyaline fungi (Visser and Whittaker, 1977). When two similar food sources were presented, the species would initially prefer one of the fungi until heavy grazing reduced its desirability. A switch then occurred to the other food source. This implied that 0. subtenuis preferred not only certain species of fungi, but that these preferred organisms had to be in an actively growing state for intensive feeding to occur and to continue. Inter- and intraspecific feeding preferences of Folsomia candida were investigated by Moore et al (1987). When actively metabolizing hyphae and senescenthyphae were offered, F. candida chose active hyphae over senescent hyphae. In another case, spores were preferred over active hyphae. Results of this study indicated that the "donor- controlled" model (Pimm, 1982) would be inadequate for certain microarthropod species as suggested by earlier studies (Macfadyen, 1963; Behan and Hill, 1978). Leonard 75 (1984) also demonstrated that F. candida preferentially grazed younger hyphae over older hyphae. Results of a study by Matic and Koledin (1985) also indicated that a collembolan species was selective in its feeding habits. When the preferred food item was offered, the species fed more heavily than when the less preferred food was offered. Different patterns of oviposition have been reported for various collembolan species. Snider (1971) found a distinct pattern in the succession of ecdyses and oviposition in Folsomia candida. This species produced eggs in every other instar after reaching maturity, and fecundity was high with an average of 1,011 eggs in a female’s lifespan. Temperature also played an important role in instar duration, mortality and egg production (Snider and Butcher, 1973). Animals kept at 15°C had the longest lifespan and highest egg production compared to individuals at higher (21°C and 26°C) temperatures. Onychiurus armatus exhibited a different egg-laying pattern, producing eggs in every instar after maturity (Snider, 1974). Individuals cultured at 15°C had a longer lifespan and higher survival rate than those kept at higher temperatures. The largest egg clumps were produced at 15°C, but fewer individuals laid eggs. As a result, cumulative average egg production was actually lower at 15°C than at 21°C. In a study of dietary influence on growth and fecundity of Onychiurus justi (Denis) = Onychiurus folsomi (Schaffer), Snider (1971) found that type and quality of the diet could affect growth, survival and fecundity of the species. The amount of protein and fat in the diet probably was responsible for observed differences. Animals fed a higher protein content food produced more eggs than those fed on lower protein 76 food. But individuals fed on a lower protein diet showed a stabilized survival rate after an initial period of decline, and mortality was lowest among all diets after 100 days. The size of individuals fluctuated after they reached a certain instar; there could be decrease in length instead of increase. Size fluctuations continued until the animals reached senility; at that point, body size decreased a small amount with each succeeding instar until death. A The nitrogen concentration in food significantly affected both the fecundity and molting rate of Collembola (Booth and Anderson, 1979). Folsomia candida was offered two fungi, each grown in liquid media containing 2, 20, 200 and 2,000 ppm nitrogen respectively. Although treatment effects differed according to fungal species, individuals feeding on both species exhibited increased rates of molting and egg production up to 200 ppm N and inhibition of growth and fecundity at 2,000 ppm N. The inhibitory effect of high N concentration could have been the result of fungal metabolic products under high nutrient conditions, an amino acid imbalance, or toxicity caused by the high concentrations of asparagine used as nitrogen source. Kurup and Prabhoo (1982) observed that Cryptopygus thermophilus (Axelson) had a higher fecundity (total egg production), larger egg clutch size, longer lifespan and reproductive period when fed on yeast, compared to three other food sources (one fungus and two species of decaying leaves). By monitoring the start of the reproductive period and weekly egg production, van Amelsvoort and Usher ( 1989) found that Folsomia candida showed different life history strategies depending on food source. When fed on a more nutritious food (baker’s yeast), life processes were faster (indicated by early maturation, early maximum egg 77 production, early loss of fertility and early death) than when fed on some less nutritious foods (leaf litter mixed with small amount of baker’s yeast). Walsh and Bolger (1990) showed that the food preferences of three collembolans, Onychiurus fiircifer (Borner) , Hypogastrura denticulata (Bagnall) and Isotomina thermophila (Axelson), differed significantly. Onychiurus fiircifer achieved greatest population size and fastest growth rate when feeding on one of its least preferred food items. They indicated that nutrient content of the diet overrode the effect of preference. Food type did not have a definite effect on population growth of H. denticulata, and this species was generally thought of as an opportunistic species and an early colonizer of many habitats. As such it would not be as selective in its foods as 0. fiircifer, which was generally not considered a pioneer species. It has been demonstrated that selective grazing by Collembola could significantly affect leaf microfungal succession in microcosms (Klironomos et al, 1992). Folsomia candida preferred primary saprophytes over secondary saprophytes. Generally, when feeding on primary saprophytes the individuals grew larger and produced more eggs than when feeding on secondary saprophytes. Colonization of litter fragments by primary saprophytes did not change very much in the absence of collembolans, although litter fragments were also colonized by secondary saprophytes. In the presence of F. candida, primary saprophytes were almost totally eliminated after the second week, allowing for increased colonization by secondary saprophytes. Total number of litter fragments colonized by fungi was higher in the presence of F. candida. Selective grazing by Collembola also altered the outcome of competition between two basidiomycetes in the field (Newell, 1984a and b). Onychiurus latus Gisin showed 78 a significant preference for one fungus over another. The preferred fungus was a more rapid colonizer, and would outcompete the other fungus in the absence of 0. latus. In the presence of the collembolan, the two fungi co-existed in the field, but occupied different layers. The preferred fungus colonized a layer that was less frequented by 0. latus. Most studies to date have concentrated on food preference of single species. Data on food preference of coexisting soil collembolan species are scarce. The effects of a given food on the life history of different species have rarely been investigated. For coexisting species, selection of different foods would reduce interspecific competitive pressure and increase niche separation. The present experiments employed feeding preference tests and life history observations to assess the following biological traits among coexisting soil collembolan species: 1). Differences in food preferences; 2). Adaptive significance of preferences for microbial foods. 3). Differences between collembolan species in terms of life history in response to diets. Materials and Methods Species All animals and microorganisms used in this study were extracted from soil of the Control site of the ELF ecological monitoring project (Snider and Snider, 1987). The site was located in Dickinson County, Upper Peninsula, Michigan. It was a secondary 79 growth deciduous forest dominated by maple and basswood (Snider and Snider, 1987). Soil was placed in Tullgren funnels for three days, and heat in the funnels was increased gradually by controlling voltage. Extracted Collembola were collected in culture jars containing a substrate of plaster of Paris and charcoal, which were placed beneath the funnels, and replaced every twelve hours. Collembola were transferred to clean culture jars. Four coexisting species were selected: Onychiurus armatus (Tullberg), T ullbergia granulata Mills, Tullbergia yosiii Rusek and Tullbergia iowensis Mills, because of their amenability to laboratory culture. All four are soil—dwelling Onychiuridae, and are parthenogenetic under laboratory conditions. Groups of ten adults were transferred to plastic culture jars provided with a 1:1 plaster of pariszcharcoal mixture (Snider et al, 1969). After eggs had been laid, the adults were killed and mounted on slides for species identification. Culture jars with eggs were kept for further use if all ten adults belonged to the same species, or were discarded if they contained two or more species. Animals were kept at room temperature (approximately 220C) and fed baker’s yeast. For isolation of soil fungi, potato dextrose agar (200 g potato, 20 g dextrose and 15 g agar per liter) was prepared. After autoclaving (120°C, 15 minutes), 50 mg streptomycin sulfate was added to 250 ml PDA before pouring into Petri dishes. A drop of dilute soil suspension was spread over the surface of the hardened PDA. Since incubators were located in another building, and transportation may have resulted in contamination, cultures were kept at room temperature (approximately 220C) in the laboratory under an isolation hood. Cultures were checked daily for microbial growth. Once growth of microorganisms had occurred, the single colony transference method was 80 employed to purify the cultures. More than 50 isolates were obtained. Six fungi: Mucor sp., Acremonium sp., Absidia sp., Humicola sp., Penicillium sp. 1, Penicillium sp. 2 and an actinomycete, were identified and selected for food preference tests. These microorganisms were kept in agar slants and were periodically subcultured, so that all foods offered to the Collembola stemmed from the same colonies. Food Preference Tests All possible pairwise combinations of the six fungi and an actinomycete were tested against each other (21 total). The microorganisms were subcultured from slants to petri dishes with PDA and allowed to grow for one week. Autoclaved (120°C, 30 minutes) plaster of Paris and charcoal were mixed with sterile distilled water and poured into 5-cm-diameter test jars, which were sterilized by immersion in 1% sodium hypochlorite (Clorox) solution and rinsed with sterile distilled water. Jars were allowed to set for three days before introducing microbial foods and animals. Food offered to the Collembola consisted of agar discs (0.5 cm diameter) with microbial growth. Each test jar received two discs (two microbial species) placed on the opposite sides of the jar perimeter. Sixteen adult animals which had been starved for 24 hours were introduced in the center of the jars, which were kept at room temperature and in the dark by a covering of black plastic film. The number of individuals feeding on each food item was recorded five times at 12 hour intervals. Five or six replicates were used in each test. Soil Collembola had a tendency to aggregate around the food and to stay there until it was exhausted. Possibly the same animals remained near a food disc over two 81 or more observation periods. Hence, the time variable was not independent over successive intervals, and was treated as repeated measurement. The number of individuals feeding on each food item was converted to percent of total animals feeding and transformed to loge [p/(l-p)] for analysis. The homogeneity of the time variable (over time, no differences in the proportion of animals feeding on a given food item) was checked first. If not significant (P > .05), then multivariate analysis of variance was performed to test for food preference. If significant differences occurred over time, then data from each observation were tested separately. Generally there were three situations in which the time variable yielded significant differences. 1) One food item was preferred, but the degrees of preference were different over time. 2) A species showed preference at certain times but not at others. If three or four (out of five) times the species showed significant preferences, existence of a true preference was assumed. 3) A species preferred one food item at certain times but preferred another at other times. Life History Studies Based on results of food preference tests, two species, 0. armatus and T. yosiii, were selected for life history studies. Mucor sp. , Acremonium sp. and the actinomycete were offered as sole food sources to these animals. There was a total of six combinations (two collembolan and three microbial species). Each combination was replicated five times. Age of experimental animals was standardized, by using those which had hatched 82 from eggs within 24 hours. Twenty newly-hatched individuals were introduced to each replicate jar (5 cm diameter, 2.5 cm height), which contained a substrate of plaster of Paris and charcoal. Microorganisms were kept in slants and were periodically subcultured in Petri dishes, where they were allowed to grow for one week. Agar discs (with microbial growth) were then cut out (0.5 cm diameter) and placed in the center of the rearing jars. Food discs were changed every three or four weeks to reduce contamination and to prevent the possibility of starvation. Rearing jars were incubated in the dark at a constant temperature of 20°C. They were checked twice a week and moistened with sterile distilled water once a week. Individuals were transferred to new jars every six weeks to control contamination. An extra jar of every combination was kept in the same conditions as the test jars. When death of individuals in test jars caused unbalanced numbers of animals, replacements from the reserve cultures were used to keep numbers of animals approximately even. Growth, survival and egg production of the Collembola were monitored. The experiment lasted until the average survival rate reached 10%. During the experiment, five living individuals were randomly selected in each jar, and their body length (including head) was measured every week in the first 25 weeks, once every two weeks thereafter. Exuviae produced by T. yosiii were counted and removed twice a week. Dead animals were recorded and removed. Eggs were counted and removed twice a week. Body length, survival and egg production were compared between three dietary treatments for 0. armatus and T. yosiii. In addition, production of exuviae by T. yosiii 83 was analyzed. All statistical tests were performed using Tukey’s test for each individual week. In the case of 0. .amzatus, the group fed on Acremonium sp. had a shorter lifespan than the others, and t—tests was employed to compare the two remaining food treatments during the last few weeks. Results A. Food Preference Each collembolan species showed different food preferences. Onychiurus armatus preferred both Mucor sp. and Absidia sp. over other food items in four out of six tests. It chose the actinomycete over both Penicillium sp.1 and sp.2, and Humicola sp.. Acremonium sp. was preferred over both Penicillium sp.2 and Humicola sp.. Penicillium sp.2 was consumed only when Penicillium sp.1 was offered as an alternative. Penicillium sp.1 and Humicola sp. were generally avoided (Table 15a). Mucor sp. , Absidia sp. and Acremonium sp. were preferred foods for T. granulata, over all other four food items. When the three readily consumed microbial foods were compared, no significant differences emerged (Table 15b). T ullbergia granulata did not select Humicola sp., Penicillium sp.2, or the actinomycete, and only fed on Penicillium sp.1 more frequently than on Humicola sp.. Compared to 0. armatus and T. granulata, both T. yosiii and T.i0wensis were less discriminating in their choice of foods. Tullbergia yosiii and T. iowensis only showed six and three significant preferences out of 21 tests, respectively. Acremonium 84 Table 15. Summaries of results of feeding preference experiments. a. Onychiums armatus X axis: Y axis: Absi. Mucor Acre. Peni.1 Peni.2 Humi. Acti. Absidia sp. = > > = > > Mucor sp. > > = > > Acremonium sp. = > > = Penicillium 1 < = < Penicillium 2 — < Humicola sp. < actinomycete b. Tullbergia granulata X axis: Y axis: Absi. Mucor Acre. Peni.1 Peni.2 Humi. Acti. Absidia sp. = = > > > > Mucor sp. = > > > > Acremonium sp. > > > > Penicillium l = > = Penicillium 2 Humicola sp. actinomycete 85 Table 15 (continued). c. Tullbergia yosiii X axis: Y axis: Absi. Mucor Acre. Peni.1 Peni.2 Humi. Acti. Absidia sp. = = = > Mucor sp. < = VV ll Acremonium sp. > > Penicillium 1 Penicillium 2 Humicola sp. = actinomycete d. Tullbergia iowensis X axis: Y axis: Absi. Mucor Acre. Peni.1 Peni.2 Humi. Acti. Absidia sp. = = > = = > Mucor sp. = = = = = Acremonium sp. = = = = Penicillium 1 < = = Penicillium 2 = = Humicola sp. = actinomycete Symbolic abbreviations are as follows: >: Y axis food is significantly preferred over X axis food. <: X axis food is significantly preferred over Y axis food. =: no significant difference between food items. 86 sp. was the preferred food item for T. yosiii, over Mucor sp. , Humicola sp, Penicillium sp.1 and 2, but not over Absidia sp. and the actinomycete. Tullbergia yosiii also preferred Mucor sp. over Humicola sp., and Absidia sp. over Penicillium 2 (Table 150). Tullbergia iowensis only showed preference for Absidia sp. over both Penicillium sp.1 and the actinomycete, and for Penicillium sp.2 over Penicillium sp.1 (Table 15d). Results suggest that under controlled conditions, T. yosiii and T. iowensis are general feeders, and T. granulata and 0. annatus are selective feeders. B. Growth Onychiurus armatus Body length of 0. armatus increased very rapidly at first, from newly hatched to fully developed in about 15 weeks (Figure 19). After this period, average body length remained unchanged or even decreased slightly, depending on the food source they were offered. When fed Mucor sp. , average body length fluctuated during this period. Fed Acremonium sp. , it decreased gradually until week 30, then increased again . After reaching maximum size, individuals fed on the actinomycete decreased in size for a few weeks and then fluctuated until week 40. During the last few weeks of their life, body length usually increased again. There were no significant size differences among animals feeding on three different food items during the first 15 weeks, with only minor, temporary exceptions (Appendix II). From week 16 until week 24, individuals feeding on the actinomycete were smaller than those feeding on both Mucor sp. and Acremonium sp.. Late in their life (from week 27 to the end of the experiment), animals feeding on Mucor sp. had 900 - :2: “Illlllllllllllln I ll 1 l l l I I I i A i i I 600- III 800 r Mucor sp. Body Length (um) 200 I I I I I I 900 r 800 - 700 r III“ 4oo~ I. 300—I Body Length (um) 200 I I I I I 1 900 7 800 a 700, H[IIHIIIIIIIIIIIIIIIIIIHIIIIIHl 600— “1 actinomycete Body Length (urn) 400 - 300‘; 200 1 , , , I 1 Week Figure 19. Body length of 0. armatus fed on three different microbial foods. Vertical lines represent standard deviations. 88 significantly larger bodies than those feeding on either Acremonium sp. or the actinomycete (Appendix H). Tullbergia yosiz'i After hatching, most individuals of T. yosiii increased in length very rapidly in the first ten weeks. Individuals fed Mucor sp. and Acremonium sp. kept increasing in size after the first ten weeks, although at a much slower pace (Figure 20). The size of individuals fed the actinomycete fluctuated during this period (Figure 20). Comparing animals feeding on the three different food items, those fed on the actinomycete usually had the shortest body length, although sometimes this was not statistically significant (Appendix III). Individuals fed on Acremonium sp. had larger bodies than those fed on Mucor sp. from week 18 to week 39, and in the last few weeks of the experiment, those fed on Mucor sp. showed significantly greater body length than individuals fed on Acremonium sp. (Appendix III). C. Survival Onychiurus armatus Survival rate of 0. armatus was significantly influenced by food type (Figure 21). Individuals fed on Mucor sp. had the highest survival rate. The survivorship curve of this group indicated a relatively constant rate of mortality independent of age. The average lifespan (based on 10% survival) was 55 weeks. The survival rate of individuals fed on Acremonium sp. was not significantly different from those offered Mucor sp. in the first 30 weeks (Appendix IV). Mortality increased after week 25 and resulted in the 900 - 800 - Body Length (um) 300 - 200 900 — 800 - Body Length (um) 800 a 200 900 - 800 ~ Body Length (um) 800 a 200 700 — 600 - 500 - 400 ‘ Mucor sp. 700 a 600 - 500 ~ 400 n A cremon/Z/m s p . 700 r 600 - 500 ~ 400 a actinomycete l l l l l l 10 20 3O 4O 5O 60 Week Figure 20. Body length of T. yosiii fed on three different microbial foods. Vertical lines represent standard deviations. 90 shortest lifespan among the three treatments (average 37 weeks). Survival rate of juvenile animals fed on the actinomycete was relatively low, significantly so when compared to the group fed on Mucor sp.. After the first five weeks, survival rate decreased very slowly and resulted in a longer lifespan (average 60 weeks). T ullbergia yosiii Differences of the survival rates among the three groups were caused mainly by different juvenile mortalities (Figure 22). The group fed on Mucor sp. had the lowest juvenile mortality and the group fed on the actinomycete experienced the highest. There was no statistical difference between the two groups fed on Mucor sp. and Acremonium sp. (Appendix V). Their average lifespans were 63 and 61 weeks respectively. When the actinomycete was offered as food source the survival rate was significantly lower than when either Mucor sp. or Acremonium sp. were used. The average lifespan of individuals fed on the actinomycete was 56 weeks. D. Egg Production Onychiurus armatus Onychiurus armatus fed on all three microbial foods began producing eggs during the third week, numbers of eggs per individual increasing gradually toward a maximum during Week 14 or 15 (Figure 23). Thereafter, mean weekly egg production decreased and became variable. The group fed on Mucor sp. reached maximum egg production in week 15, with an average of 10 eggs per individual. From week 20 to 44, mean number of eggs 91 _..—— Mucor Sp. - - - - Acre/770mm sp. ....... actinomycete Survival Rate 00 l I I I I J O 10 20 3O 4O 5O 60 Week Figure 21. Mean survival rates of 0. armatus fed on three different microbial foods. Vertical lines represent standard deviations. 92 _.— Mucor Sp. - - - — Acre/770mm sp. ....... actinomycete Survival Rate 0.0 L l I I I I I O 10 20 3O 4O 50 6O 70 Week Figure 22. Mean survival rates of T. yosiii fed on three different microbial foods. Vertical lines represent standard deviations. 93 fluctuated between 2 and 5, and after week 45 fewer than 2 eggs were produced per week. This group showed a higher cumulative egg production than the other two groups during its entire lifespan (Figure 24), on average producing a total of more than 180 eggs per individual. After reaching maximum egg production in week 14, individuals fed on Acremonium sp. gradually decreased their egg-laying activity (Figure 23). After week 27, this group usually did not produce any eggs at all. Cumulative egg production was generally higher than that of animals fed on the actinomycete, but only from week 8 to 22 (Figure 24, Appendix VI). Average total production per individual was 73 eggs in 37 weeks. Individuals fed on the actinomycete showed a relatively constant weekly egg production from weeks 10 to 26. Egg production was highest in week 14 and was the lowest (about 5 per individual) among the three groups (Figure 23). Total egg production, however, surpassed that of animals fed Acremonium sp. (Figure 24). On average, 108 eggs were laid over an individual’s lifespan. Tullbergia yosiii Individuals fed on both Mucor sp. and Acremonium sp. began laying eggs in the third week, whereas those fed on the actinomycete began in the fourth week (Figure 25). All three groups reached maximum egg production in week 14. Overall, egg production among these three groups was very similar. Analysis of cumulative egg production data (Figure 26) yielded very few significant differences (Appendix VH). In all three groups, an average of approximately 280 eggs per individual were laid over 63 weeks. 94 258 33838 “5st 8:: so 3% 39:5 .0 3:232: Ba 5:03on wwo 3603 :82 .mm oSwE O I N I? I I © 1 do \Qbo§\ @ I w ooooficocpom I I do SQQQSmka _U I Or M fi~.________2___~___fl2_2__2_._Nrw .6 00 ON 9N kw ON ON QN nu «N .N ON or 9 t or or «r or N— _.r 0.. o 0 b 0 0 1 0 N 3 RU RU 3 do \Qb\o§\ oeoo>Eoczom _Q m SQQQRxmkb T NOGDCOYTNO T'— 95 200 — 150 — (I) O) 8’ z "Ln-T g 100 — " “_6 3 E 3 o 50 - _Mucor sp. — - - - Acre/770mm sp. ....... actinomycete O I l I I l O 10 20 30 4O 5O 60 Week Figure 24. Mean cumulative egg production per individual 0. armatus fed on three different microbial foods. Vertical lines represent standard deviations. 96 .€o£ 33828 Baotou 8:: so we“ 38% H 3:232: Ba 5503on wwo >283 532 .mm 835 v_®®>> 00 NO r 0 00 00 00 no 00 00 10 00 N0 r0 on 0‘ D! A. 0* 0' vv 9 mt rt 01 on 00 ha 00 DO 1n 00 do \be§\ @ oeoo>Eoczom I .Q m SQQQRQKQ T _U (\JOCDCOVNO v.7. 9663 N ueeIAI NM r0 00 ON 06 NN ON ON {a rd ON 0.. Gr up Or 0— 1r Mr N». Z. 0.. O 0 b O 0 v m 'Q'A'A'Q'A'IAVA'AVA'A'A'A’A'A‘ 'A'A'A'A"'A'A'A'. V.V.'.V.V.'.'.V.'.V.'.V.'.'.'.'.'.'.Q.v.Op, I7A'A'A'A'A'Q'A'A'A'A'A'A'A'A'A‘A'A'A'A'L'A' 'A'A" I do \Qb\§\ E 9oo>Eoczow I .Q m SQQQSmKo T D . _______fi__a___4_____fi7________ T— I NOOOLOVNO ‘— 97 300 — .'....1. 1 .1 8, 200 - , III 1 Z ,/ {Isl .02) 1' if]. I; {,1’ _.. Mucor sp. 3’ 100 _ , f... - - - - ACmeON/Z/m Sp. 9’1 ....... actinomycete , I 1.,” O . a: 1]" l l l L l l O 10 20 3O 4O 50 6O 70 WEEK Figure 26. Mean cumulative egg production per individual T. yosiii fed on three different microbial food. Vertical lines represent standard deviations. 98 E. Production of Exuviae by T. yosiii Cumulative exuviae production of T. yosiii (a measure of molting frequency) fed on three different foods is illustrated in Figure 27 . The animals produced exuviae from hatching to death at a relatively constant rate. They molted approximately 47 times in 63 weeks. There were no significant differences among the three food treatments, except from week 6 to 11 when feeding on Mucor sp. resulted in higher cumulative exuviae production than feeding on the actinomycete (Appendix VIII). Discussion Food preference tests showed that coexisting collembolan species have different food preferences. Results agreed fully with previous work dealing with two or more collembolan species (Singh, 1969; Walsh and Bolger, 1990). Based on their food selectivity, the species in this study could be separated into two groups: selective and general feeders. Onychiurus armatus and T. granulata were selective feeders, showing very strong food preferences when offered a choice. Tullbergia yosiii and T. iowensis were general feeders, not discriminating between foods in most of the tests. However, the two selective feeders showed differences in food selection. Both preferred Mucor sp. and Absidia sp. , but 0. armatus avoided Acremonium sp. , while T. granulata accepted it to the same degree as the other two preferred microorganisms. Onychiurus armatus preferred the actinomycete over the other three food items, while T. granulata showed 99 50 - l g 40 _ I ’ Z), I X o “J I. “5 m 30 - x1 5 x .0 § f) 20 _ . ———MUCO/ SD. f, v“ - - — — Acremon/Z/m SD. 3 4' ....... actinomycete 3 ,. O 10 - f" 0 d: I I I I I I I O 10 20 3O 4O 50 60 70 WEEK Figure 27 . Mean cumulative exuviae production per individual T. yosz'ii fed on three different microbial foods. Vertical lines represent standard deviations. 100 no preference for it. Tullbergia granulata exhibited a striking pattern in its selection of foods, grouping them into either preferred or not preferred foods. Given a choice between the two groups, it always clearly preferred one; given a choice within each group, it did not show any selectivity in most of the tests. Although differentmicroorganisms were used for food, results for 0. armatus were comparable to those of Shaw (1988). In his study, 0. armatus showed significant preferences in 43 tests out of 66 performed. In the present investigation, 0. armatus showed 14 significant preferences in 21 tests. In both cases, approximately 2/3 of the tests yielded significant results. When statistics are performed to test the significance of feeding preferences, one must be very careful in how to deal with the time variable. Because of the aggregative nature of these animals, the assumption of independence of successive observations is questionable. Erroneous assumptions of statistical independence lead to spuriously high degrees of freedom, a condition termed "pseudoreplication" which permeates previous work on collembolan feeding preferences. Ideally each "observation" should be a separate experiment, but this would require an incredible amount of labor and resources. The method used here treated the time variable as a repeated measurement and reduced its degrees of freedom. Results are, thus, much more conservative as well as robust. Snider (1971) stated that after a certain instar the size of 0. folsomi fluctuated and bore no relation to the instar they were in. Isotoma viridis Bourlet lost weight, or live weight remained steady, during the period of senile molting (Zettel, 1982). Variable body length was also observed in the present experiment. Previous studies have used body length or biomass to predict population growth. If age or instar of these animals 101 were unknown, a serious problem could result. Considering 0. armatus feeding on Acremonium sp., in week 31 its body length was almost the same as in week 10. In week 31 the animals were senile and had lost almost all of their fertility; in week 10, however, individuals were just beginning to reach maximum reproductive potential. The contribution of these two age groups to population growth would thus differ considerably, despite similar body size. According to Snider (1974), 0. armatus reached 10% survival rate between 200 and 420 days, and produced 98 to 124 eggs in its lifespan at temperatures ranging from 150C to 260C. These results were very similar to those obtained here. The lifespan of individuals fed on three different foods ranged from 37 to 60 weeks, and cumulative egg production ranged from 73 to 182. Food preference and life history data agreed perfectly in the case of 0. armatus. The species favored Mucor sp. over the other two microbial foods. Fed on Mucor sp., it showed a much higher egg production than when reared on the other two foods. Its early survival rate was higher than when fed the actinomycete, and its lifespan was longer than when fed Acremonium sp.. Results thus suggested that 0. armatus preferred the food that was best for its life processes. Comparing Acremonium sp. to the actinomycete (both of which had equal preference ratings), 0. armatus showed better early survival and higher egg production when fed on Acremonium sp., but a longer lifespan when fed the actinomycete. These data can be interpreted in terms of life history: higher survival and egg production early in life compensate for the shorter lifespan. A diet can be termed high quality when the cultured animals show the highest reproductive rate and the lowest mortality (Snider, 1971). Therefore, Mucor sp. can be 102 considered a high quality food for 0. armatus. Although 0. armatus was identified as a selective feeder, in the broader sense the species can survive and reproduce well on a wide range of food items. Present data conflict, however, with results obtained by Walsh and Bolger (1990) in that 0. furcifer achieved largest population size and fastest growth rates when fed on one of its least preferred foods. It is not clearly understood why T. yosiii selected Acremonium sp. over Mucor sp. and other microbial foods in preference tests. There were no significant differences in survival, egg production and exuviae production when either Acremonium sp. or Mucor sp. were offered as food. Visser and Whittaker ( 1977), Addison and Parkinson (1978), Shaw (1988), Rusek (1989) and Schultz (1991) suggested that the existence of toxic materials resulted in avoidance behavior, which could explain avoidance of Penicillium sp.1 and 2 by T. yosiii. Mills and Sinha (1971) found that H. tullbergi liked fungi with low mats but both Acremonium sp. and Mucor sp. used here had similar low mats. Bengtsson et al (1988) observed that odors of fungi played an important role in attracting Collembola. Schultz (1991) found that hyphal diameter and cell wall thickness bore no relationship to food selection by Collembola. It can only be speculated that some other physical or chemical characteristics of the two food items may have induced the species to favor Acremonium sp. over Mucor sp.. From a nutritional point of view, these two food items did not cause any differences in growth, survival, egg and exuviae production. The only disadvantage when T. yosiii fed the actinomycete was the higher juvenile mortality. Once individuals survived to the adult stage, the actinomycete provided almost 103 the same nutritional value as the other two foods to adults in terms of survival rate, egg production and molting frequency. It is concluded that, as a general feeder, T. yosiii can grow equally well on different foods. Under natural conditions, food is usually a major limiting factor to population growth. With the ability to survive well over a wide range of diets, T. yosiii could reduce competition by feeding on foods which are not preferred by other animals. van Amelsvoort and Usher (1989) found that Folsomia candida adopted a " fuzzy" life history strategy along the r-K spectrum. It appeared to be able to alter its position in response to feeding conditions: r-selected when conditions were good, but apparently K-selected when conditions were poor. The theory is only relevant to 0. armatus feeding on Acremonium sp. and the actinomycete in this study. According to their theory, Acremonium sp. should be considered a higher quality food for 0. armatus than the actinomycete, because feeding on Acremonium sp. speeded up life processes by means of higher early egg production and a shorter lifespan. Onychiurus armatus and T. yosiii showed totally different preferences and life histories on different food sources. It is of ecological importance for coexisting species to separate their resources. The different adaptations of these collembolans demonstrated that they had different food preferences, and these different behaviors were clearly reflected in their life processes. Soil is a heterogeneous environment. The distribution of soil animals is influenced principally by habitat characteristics (e. g., soil moisture, temperature, pore space, chemistry). Often these parameters do not explain either distinct niche boundaries or overlap (Moore et al, 1987). Soil Collembola exhibit the ability to utilize a wide 104 range of food sources, different food preferences, and variable life processes resulting from them. All of these characteristics contribute to niche separation of coexisting species, representing the boundaries that limit the degree of overlap between soil Collembola with similar abiotic requirements. SUNIMARY I. Mouthpart Structures and Gut Contents Study 1. Three coexisting collembolan species, Isotoma notabilis Schaeffer, Sminthurinus henshawi (Folsom) and Orchesella hexfasciata Harvey, were collected from two hardwood forest sites in Michigan’s Upper Peninsula. 2. Scanning electron microscopy showed that the three species had very similar mandible structures. They were all of the "biting" type, consisting of a dental portion and a fulcrum. The dental portion included apical teeth and molar plates. The apical teeth of the right mandibles varied in number from four to six, depending on the species, and also showed intraspecific variation. The left mandibles invariably had four apical teeth. There were some small differences between species in molar plate structure. The right molar plate had a projection at the end, and the left molar plate had a depression at the end which was not mentioned in any previous studies. The functions of molar plates were briefly discussed. 3. Mandible length was measured. There was very little overlap between species. Isotoma notabilis had the shortest mandibles, 0. hexfasciata had the longest, whereas those of S. henshawi were intermediate. 4. All three species had maxillar heads consisting of a three-toothed ungulum and some lamellae. The number and shape of lamellae differed between species. Isotoma notabilis had six lamellae, and lamella l was long and branched to form an apical rake. There were only five lamellae in S. henshawi and 0. hexfasciata, and their shapes were 105 106 different in each species. 5. The general structure of the labrum was very similar in all species, except for chaetotaxy. A regression model was built to estimate labrum width based on head, trunk and mandible lengths (R2 > 0.95). There was very little overlap in labrum width between these species. 6. Numbers of individuals with gut contents were counted, and food components in the intestines were examined. Interspecific comparisons were based on 1988 data for the three species, and intraspecific between-year analyses were based on 0. hexfasciata specimens obtained in 1988 and 1990. 7. Proportions of individuals with gut contents were the result of species-specific feeding behavior and environmental conditions. Long periods of drought depressed food availability for these species. Seasonal variations in proportions with gut contents were found in all three species. There were few differences between the two sites. Orchesella hexflzsciata showed some variation between years, differences being caused by the long period of drought in 1988. 8. The three collembolan species consumed different particle sizes, corresponding well with the size of their mouthparts. Isotoma notabilis had the smallest mandibles and labrum, and fed mainly on small and medium size particles. Sminthurinus henshawi’s mandibles and labrum were intermediate, and the species fed on approximately equal amounts of small, medium and large particles. Orchesella hexfasciata had the largest mandibles and labrum, and it ingested mostly large particles. 9. Eight food categories were recognized in the animals’ guts. There were distinct differences between the three species. Fungal hyphae and colloidal materials 107 were the main components in the diet of I. notabilis. A greater proportion of fungal spores was found in S. henshawi. More diversified foods were encountered in 0. haxfasciata. Seasonal variations were the results of food availability. There were few between-site differences in all three species, and few between-year variations in 0. hafasciata. 10. Results showed that coexisting species utilized different food types and particle sizes, thereby reducing interspecific competition. 11. Food Preference and Life History Studies. 1. Food preference was studied in four soil-dwelling Collembola, Onychiurus armatus (Tullberg), Tullbergia granulata Mills, Tullbergia yosiii Rusek and Tullbergia iowensis Mills. Seven microbial taxa were offered as foods. Onychiurus armatus and T. granulata were selective feeders, showing significant preferences. Tullbergia yosiz’i and T. iowensis were general feeders, less discriminating in their choice of foods. There were also interspecific differences in preference within each selective or general feeder species group. 2. Growth, survival and egg production were monitored for 0. armatus and T. yosiii fed on three different microbial foods. Production of exuviae by T. yosiii was also observed. 3. After an initial growth period, average body length of both 0. armatus and T. yosiii remained unchanged or even decreased, depending on the food source they were offered. It was found that body size bore no relation to reproductive potential, and the usage of body mass or size to predict population growth was questioned. 108 4. When fed on the most preferred food, Mucor sp. , 0. armatus achieved highest egg production per individual (more than 180), and a fairly long lifespan (average 55 weeks based on 10% survival). Feeding on Acremonium sp. resulted in better early survival and higher egg production than feeding on the actinomycete, but the latter diet induced a much longer lifespan (60 weeks). 5. There were very few variations in egg and exuviae production per individual when T. yosiiz' was reared on three different microbial foods. Feeding on the actinomycete resulted in an higher juvenile mortality, but it provided the same nutritional value to adult T. yosiii as the other two microbial foods when measured by survival, egg production and molting frequency. 6. Studies showed that these species had the ability to use a wide range of foods and exhibited different food preferences, and that diet affected the animals’ life processes. These observations document mechanisms of niche separation between coexisting species; these mechanisms tend to limit the degree of overlap between soil Collembola with similar abiotic requirements. IMPROVEMENT AND FURTHER STUDY This study documented food partitioning mechanisms of coexisting Collembola, based on selection of different food types and particle sizes, as well as on preferences for microbial foods. Additional field and laboratory studies could further define the precise mechanisms of niche separation in Collembola. The ecology of coexisting collembolans under field conditions needs to be studied in detail. Indirect evidence in the present study suggested that species have different vertical and horizontal distribution patterns. Thorough investigation of distribution patterns on a small scale would help define the microhabitats these animals occupy. Data on food availability in the field could yield information to explain seasonal variation in food consumption. Distribution of pollen and fungal spores, for instance, is not only seasonal, but also heterogeneous vertically and horizontally. Combined with the distribution patterns of animals, these data could prove food availability to be an important factor in collembolan feeding habits. Food preferences of several species known to coexist in the field in quantifiable microhabitats (e.g. , litter, A horizon, or rhizosphere) need to be more fully documented. Stock cultures of these animals need to be established. Different fungal hyphae and spores (cultured from gut contents rather than from soil), pollen of known origin, plant debris (with and without microbes) could be offered to Collembola as choices. Most importantly, correlation between field-derived and laboratory-derived information is still scarce, partly due to lack of appropriate methods. These include 109 110 rearing techniques (efforts to rear 0. hexfasciata in this study, for instance, were unsuccessful) as well as tools for isolating, culturing and identifying microbial taxa. My results indicate that a combination of data on collembolan anatomy, feeding preferences, and microhabitat characteristics (these need to be fully quantified) would yield reliable information on the functional position of Collembola in the soil/litter system. LIST OF REFERENCES LIST OF REFERENCES Addison, J. A. and D. Parkinson. 1978. Influence of collembolan feeding on soil metabolism at a high arctic site. Oikos, 30: 529-538. Agrell, I. 1940. A study of the collembolan population in the Tometask region of northern Lappland, including overwintering conditions and migrations in the field at different periods. Ent. Medd. 22: 12-13. Amelsvoort, van P. A. M. and M. B. Usher. 1989. Egg production related to food quality in Folsomia candida (Collembola: Isotomidae): effects on life history strategies. Pedobiologia, 33: 61-66. Anderson, J. M. and I. N. Healey. 1972. Seasonal and inter-specific variation in major components of the gut contents of some woodland Collembola. J. Anim. Ecol., 41:359-368. Barnes, B. V. and W. H. Wagner, Jr. 1981. Michigan trees. Univ. Mich. Press, Ann Arbor, Mi., 384pp. Behan, V. M. and S. B. Hill. 1978. Feeding habits and spore dispersal of oribatid mites in the North American Arctic. Rev. Ecol. Biol. Sol, 15: 497-516. Bengtsson, G., A. Erlandsson and S. Rundgren. 1988. Fungal odor attracts soil Collembola. Soil Biol. Biochem., 20: 25-30. Bockemiihl, J. 1956. Die Apterygoten des Spitzberges, eine faunistisch-okologische Untersuchung. Zool. 1b., Syst., Jena, 84: 113-194. Bodvarsson, H. 1970. Alimentary studies of seven common soil inhabiting Collembola of southern Sweden. Ent. Scand., 1: 74-80. Booth, R. G. and J. M. Anderson. 1979. The influence of fungal food quality on the growth and fecundity of Folsomia candida (Collembola: Isotomidae). Oecologia (Berl.), 38: 317-323. 111 112 Borner, C. 1908. Collembolen aus Siidafrika nebst einer Studie fiber die 1. Maxille der Collembolen. In: Schultze’s Forschungsreise im westlichen und zentralen Siidafrika 1903-1905. Denkschr. med.-naturw. Ges. Jena, 13: 53-68. Borror, D. J., D. M. DeLong and C. A. Triplehom. 1976. An introduction to the study of insects. Holt, Rinehart and Wilson, New York, p. 738. Christiansen, K. 1964. Bionomics of Collembola. Ann. Rev. Ent., 9: 147-178. Collins, N. J., J. H. Baker & P. J. Tilbrook. 1975. Signy Island, Maritime Antarctic. In: Rosswall, T. & O. W. Heal (eds.), Structure and function of tundra ecosystems. Ecol. Bull. (Stockholm), 20: 345-374. Cummins, K. W. 1973. Trophic relations of aquatic insects. Ann. Rev. Entomol., 18: 183-206. Fjellberg, A. 1984. Maxillary structures in Hypogastruridae (Collembola). Ann. Soc. r. 2001. Belg. 114: 89-99. Fjellberg, A. 1985. Recent advances and future needs in the study of Collembola biology and systematics. Quaest. Ent., 21: 559-570. Folsom, J. W. 1899. The anatomy and physiology of the mouth-parts of the collembolan, Orchesella cincta L. Bull. Mus. Comp. Zool. Harv. 35: 7-39. Folsom, J. W. 1900. The development of the mouth-parts of Anurida maritima Guer. Bull. Mus. Comp. Zool. Harv., 36: 87-157. Gilmore, S. K. and E. M. Raffensperger. 1970. Foods ingested by Tomocerus spp (Collembola, Entomobryidae), in relation to habitat. Pedobiologia, 10: 135-140. Goto, H. E. 1972. On the structure and function of the mouthparts of the soil- inhabiting collembolan Folsomia candida. Biol. J. Linn. Soc., 4: 147-168. Harbach, R. E. 1977. Comparative and functional morphology of the mandibles of some fourth stage mosquito larvae (Diptera: Culicidae). Zoomorphologie, 87: 217-236. Hoffmann, R. W. 1905. Uber die Mundwerkzeuge und die Funktion der Kauwerkzeuge von Tomocerus plumbeus L. H. Beitrag zur Kenntnis der Collembolen. Z. wiss. Zool. 82: 638-663. Hoffmann, R. W. 1908. Uber die Morphologie und die Funktion der Kauwerkzeuge und iiber das Kopfnervensystem von Tomocerus plumbeus L. III. Beitrag zur Kenntnis der Collembolen. Z. wiss. Zool., 89: 598-689. 113 Hudson, H. 1968. The ecology of fungi on plant remains above the soil. New Phytologist, 67: 837-874. Ingold, C. T. 1971. Fungal spores: their liberation and dispersal. Oxford Univ. Press, Oxford, 302pp. Joosse, E. N. G. and E. Veltkamp. 1970. Some aspects of growth, molting and reproduction in five species of surface dwelling Collembola. Neth. J. Zool. , 20: 315-328. Joosse, E. N. G. and E. J. Testerink. 1977. The role of food in the population dynamics of Orchesella cincta (Linné) (Collembola). Oecologia 29: 189-204. Kaneko, N. 1988. Feeding habits and cheliceral size of oribatid mites in cool temperate forest soils in Japan. Rev. Ecol. Biol. Sol, 25: 353-363. Kendrick, W. B. and A. Burges. 1962. Biological aspects of the decay of Pinus sylvestris leaf litter. Nova Hedwigia, 4: 313-342. Kevan, P. G. and D. K. McE. Kevan. 1970. Collembola as pollen feeders and flower visitors with observations from the high Arctic. Quaest. Entomol. , 6: 311-326. Klironomos, J. N ., P. Widden and I. Deslandes. 1992. Feeding preferences of the collembolan Folsomia candida in relation to microfungal successions on decaying litter. Soil Biol. Biochem., 24: 685-692. Knight, C. B. and R. A. Angel. 1967. A preliminary study of the dietary requirements of Tomocerus (Collembola). Amer. Midland Natur., 77: 510-516. Kurup, N. C. and N. R. Prabhoo. 1982. Influence of food quality on reproduction and longevity of Crytopygus thermophilus (Isotomidae: Collembola). Entomon, 7: 481-485. Leonard, M. A. 1984. Observations on the influence of culture conditions on the fungal feeding preferences of Folsomia candida (Collembola: Isotomidae). Pedobiologia, 26: 361-367. Lubliner-Mianowska, K. 1956. Investigations on the chemical composition of the pollen 1. The study of the pollen as bee food. Act. Soc. Bot. Polon., 25 : 409— 419. Lunden, R. 1954. A short introduction to the literature on pollen chemistry, Svensk. Kem. Tidskr., 66: 201-213. 114 Macfadyen, A. 1963. The contribution of the microfauna to total soil metabolism. In: Doeksen, J. and J. van der Drift (eds.). Soil organisms. North Holland, Amsterdam, pp 3-17. MacNamara, C. 1924. The food of Collembola. Canad. Ent., 56: 99-105. Manton, S. M. 1964. Mandibular mechanisms and the evolution of arthropods. Phil. Trans. R. Soc., 247: 1-183. Manton, S. M. 1977. The Arthropoda: Habits, functional morphology and evolution. Oxford Univ. Press, Oxford, 527pp. Matic, R. and D. Koldin. 1985. Preference and feeding specificity of Tetrodontophora bielanensis (Collembola, Insecta) under laboratory conditions. Rev. Ecol. Biol. Sol, 22: 121-129. McMillan, J. H. 1975. Interspecific and seasonal analyses of the gut contents of three Collembola (Family Onychiuridae). Rev. Ecol. Biol. Sol, 12: 449-457. McMillan, J. H. 1976. Laboratory observation on the food preference of Onychiurus armatus ('Tullb.) Gisin (Collembola, Family Onychiuridae). Rev. Ecol. Biol. Sol, 13: 353-364. Mills, J. T. and R. N. Sinha. 1971. Interactions between a springtail, Hypogastrura tullbergi, and soil-bome fungi. J. Economic Entomol., 64: 398-401. Moore, J. C., E. R. Ingham and D. C. Coleman. 1987. Inter- and intraspecific feeding selectivity of Folsomia candida (Willem) (Collembola, Isotomidae) on fungi. Biol. Fertil. Soils, 5: 6-12. Moore, J. C., D. E. Walter and H. W. Hunt. 1988. Arthropodregulation of micro- and mesobiota in below-ground detrital food webs. Ann. Rev. Entomol., 33: 419-439. Muraleedharan, V. and N. R. Prabhoo. 1978. Observations on the feeding habits of some soil Collembola from an abandoned field in Kerala. Entomon, 3: 207-213. Murphy, P. W. and C. C. Doncaster. 1957. A culture method for soil meiofauna and its application to the study of nematode predators. Nematologica, 2: 202-214. N ewell, K. 1984a. Interaction between two decomposer basidiomycetes and a collembolan under Sitka spruce: distribution, abundance and selective grazing. Soil Biol. Biochem., 16: 227-233. 115 Newell, K. 1984b. Interaction between two decomposer basidiomycetes and a collembolan under Sitka spruce: grazing and its potential effects on fungal distribution and litter decomposition. Soil Biol. Biochem., 16: 235-239. Petersen, H. 1982. 4. Structure and size of soil animal populations. In: Petersen, H. (ed.). Quantitative ecology of microfungi and animals in soil and litter. Oikos, 39: 306-329. Pimm, S. L. 1982. Food webs. Chapman and Hall, New York. 219pp. Poole, T. B. 1959. Studies on the food of Collembola in a douglas fir plantation. Proc. 2001. Soc. London, 132: 71-82. Poole, T. B. 1961. An ecological study of the Collembola in a coniferous forest soil. Pedobiologia, 1: 113-137. Rastin, N. , G. Schlechte, A. Huttermann and K. Rosenplanter. 1990. Seasonal fluctuation of some biological and biochemical soil factors and their dependence on certain soil factors on the upper and lower slope of a spruce forest. Soil Biol. Biochem., 22: 1049-1061. Rusek, J. 1989. Ecology of Collembola. In: Dallai, R. (ed.). 3rd International Seminar on Apterygota. University of Siena, Italy. pp 271-281. Sadaka, N. and N. Poinsot—Balaguer. 1987a. Relations trophiques entre le collembole Onychiurus zrchokkei Handschin et les feuilles de chene vert Quercus ilex L. , a divers stades de decomposition. Rev. Ecol. Biol. Sol, 24: 329-340. Sadaka, N. and N. Poinsot-Balaguer. 1987b. Determination of amino acids in leaves of evergreen oak at four different stages of decomposition. Biol. Fert. Soils, 5: 158-163. Sadaka, N. and N. Poinsot—Balaguer. 1989. Relations trophiques feuilles de chene vert (Quercus ilcx L.) - Collemboles. Influence de la qualité du materiel foliaire sur la croissance pondérale d’ Onychiurus zschokkei Handschin. Rev. Ecol. Biol. Sol, 26: 197-204. Sadaka, N ., N. Poinsot-Balaguer and J. Talin. 1988. Relations trophiques feuilles de chene vert (Quercus ilex L.) - Collemboles. Influence de la qualité du materiel foliaire sur la biologic d’Onychiurus zschokkei Handschin et Folsomia candida Willm. Vie et milieu, 39: 33-39. Schaller, F. 1949. Biologische Beobachtungen an humusbildenden Bodentieren insbesondere an Collembolen. Zool. Jb. Abt. Syst., 78: 506-525. 116 Schultz, P. A. 1991. Grazing preferences of two collembolan species, Folsomia candida and Proisotoma minuta, for ectomycorrhizal fungi. Pedobiologia, 35: 313-325. Scott, H. G. and C. J. Stojanovich. 1963. Digestion of juniper pollen by Collembola. Florida Entomologist, 46: 189-191. Seastedt, T. R. 1984. The role of microarthropods in decomposition and mineralization processes. Ann. Rev. Entomol., 29: 25-46. Shaw, P. J. A. 1988. A consistent hierarchy in the fungal feeding preferences of the Collembola Onychiurus armatus. Pedobiologia, 31: 179-187. Singh, S. B. 1969. Preliminary observations on the food preference of certain Collembola (Insecta). Rev. Ecol. Biol. Sol, 6: 461-467. Snider, R. M. 1971. Laboratory observations on the biology of Folsomia candida (Willem) (Collembola: Isotomidae). Rev. Ecol. Biol. Sol, 10: 103-124. Snider, R. M. and J. W. Butcher. 1973. The life history of Folsomia candida (Willem) (Collembola: Isotomidae) relative to temperature. The Great Lakes Entomol., 6: 97-106. Snider, R. M. 1974. The life cycle relative to temperature of Protaphorura armatus (Tullberg) (Collembola: Onychiuridae), a parthenogenetic species. The Great Lakes Entomol., 7: 9-15. Snider, R. J ., J. Shaddy and J. W. Butcher. 1969. Culture techniques for rearing soil arthropods. Mich. Entomol., 1: 357-362. Snider, R. J. 1971. Dietary influence on the growth and fecundity of Onychiurus justi (Denis) (Onychiuridae: Collembola). Ann. Zool., 225-234. Snider, R. J. and R. M. Snider. 1987. ELF ecological monitoring in Michigan. 1. Description of sites for soil biological studies. Pedobiologia, 30: 241-250. Snider, R. J. and R. M. Snider. 1989. ELF communications system ecological monitoring program: Arthropoda and earthworms. In: Compilation of 1988 annual reports of the Navy ELF communications system ecological monitoring program. Technical report E06595-6. Contract No. N00039-88-C-0065, Vol.2. 72 pp. Snider, R. J. and R. M. Snider. 1990. ELF communications system ecological monitoring program: Arthropoda and earthworms. In: Compilation of 1989 annual reports of the Navy ELF communications system ecological monitoring 117 program. Technical report E06620-4. Contract No. N00039-88-C-0065, Vol.2. 108 pp. Snider, R. J. and R. M. Snider. 1991. ELF communications system ecological monitoring program: Arthropoda and earthworms. In: Compilation of 1990 annual reports of the Navy ELF communications system ecological monitoring program. Technical report E06628-4. Contract No. N00039-88-C-0065, Vol.2. 128 pp. Snider, R. J. and R. M. Snider. 1992. ELF communications system ecological monitoring program: Arthropoda and earthworms. In: Compilation of 1991 annual reports of the Navy ELF communications system ecological monitoring program. Technical report E06200—2. Contract No. N00039-88-C-0065, Vol.2. 65 pp. Surtees, G. 1959. Functional and morphological adaptations of the larval mouthparts in the sub-family Culicinae (Diptera) with a review of some related studies by Montschadsky. Proc. R. Ent. Soc. Lond., (A) 34: 7-16. Takeda, H. and T. Ichimura. 1983. Feeding attributes of four species of Collembola in a pine forest soil. Pedobiologia, 25 : 373-381. Thibaud, J. M. 1968. Cycle du tube digestif lors l’intermue chez les Hypogastruridae (Collembola) épigés et cavemicoles. Rev. Ecol. Biol. Sol, 5: 647-655. Visser, S. and J. B. Whittaker. 1977. Feeding preferences for certain litter fungi by Onychiurus subtenuis (Collembola). Oikos, 29: 320-325. Wallace, J. B. 1975. Food partitioning in net-spinning Trichoptera larvae: Hydropsyche venularis, Cheumatopsyche etrona, and Macronema zebratum (Hydropsychidae). Ann. Entomol. Soc. Am., 68: 463-472. Wallace, J. B. and R. W. Merritt. 1980. Filter-feeding ecology of aquatic insects. Ann. Rev. Entomol., 25: 103-132. Wallwork, J. A. 1972. Distribution patterns and population dynamics of micro- arthropods of a desert soil in Southern California. J. Anim. Ecol. , 41: 291-310. Walsh, M. I. and T. Bolger. 1990. Effects of diet on the growth and reproduction of some Collembola in laboratory cultures. Pedobiologia, 34: 161-171. Walter, D. E., D. T. Kaplan and T. A. Permar. 1991. Missing links: a review of methods used to estimate trophic links in soil food webs. Agric. Ecosystems Environ., 34: 399-405. 118 Widden, P. and D. Parkinson. 1973. Fungi from Canadian coniferous forest soils. Canadian Journal of Botany, 51: 2275-2290. With, de N. D. and E. N. G. Joosse. 1971. The ecological effects of molting in Collembola. Rev. Ecol. Biol. Sol, 8: 111-117. Wolter, H. 1963. Vergleichende Untersuchungen zur Anatomie und Funktionsmorphologie der stechend-saugenden Mundwerkzeuge der Collembolen. Zool. Jb. Anat. Jena, 81: 27-100. Zettel, J. 1982. Dietary influence on the biology and metabolism of Isotoma viridis (Collembola). J. Insect Physiol., 28: 453-458. APPENDICES 119 ea. mad god 085 E .o ME; 4: .o a; weed- 388$ .e 03. Sod Sod 83 Sad Sod $3 89m $3. defines .m 48. :3. on; 805 085 Sod E .o 3: meg- 338: .~ mm 85.3 .0580 85.5 .0580 85.3 .0880 8.5.3 .0580 38% fimcfl MES H mx £32 93: H ax fiwcfl 03:25: N F 523 ease I a $9 + $9 + £5 + on n .A 880% some do 85.53 05 me 523 2: massage c8 8&8on 032:2 .H 520%? Ame. v 8 mooceotmu 38¢?me 8865 8:38 a 5 £632 “:80th .Go. N 3 concourse EweflEwa o: 88:3: 5:28 a 5 E032 25% 2F .38 m unmask was: 858$:me 9865 coufifie Equsm 05 Sam 98:3 ._. 120 £35 83? «23$ @8an 5&an 623$ £339 flaming. EH5 Baggage“ £3.85 fiance. sang €53? amide: 835 £34: «SHE “$3.3 §§§§ «SHE. £an «SHE £93? £33 833 mafia.» unease £3.48 8%: mm R mm x mm R : om a «were #83 fiflfi £93? 983$ «:13? «SHE «8an amino? 83mg 5H0: 383838 338» “Rice «SHE ammnmi £33» £33 £33 «Ram: «mange asseseew «3.4%? «SHE 8385 «SHE «Seam. «NEE «Sate emng «Snoop 82.: 2 S 2 2 E 2 S : S magi n83 «Gnome “inane 98238 «8.38 we“? a? “mom “8 HQ. “$438 «Enema 200588“ awmnmwe fisfloe Danae «fies £380 £2“an «833 anion «Enema §N§§ew «Gnome 3365 83:3 $380 53% 623% «enema 334mm manna .523 a w e e m 4 m m a $5 M83 .meoow 33838 380:6 8:: :o Bea 2332.5 .0 Lo 5m H 888 .85 fiwcfl boon >383 d S3333. .mREEm «858598 We 58w 9 26 03mm?“ 8: 3:88am“ 3: .30. v 8 mooceotnv 885:me 88ch 5:28 a E 332 880:5 Ame. N 8 ooaouotmc 88¢?me 0: 38:2: 8:28 a S ESE 08am on... .Aozmmgm 803 Saw .8 38 2.: Eco .5 H83 5595 8 38 93229 mi? oocwouEwu 8865 agar/2V 285% 05 Bad €033 * 121 Sfimfl SHE. 29%th nmmfimfi nwmfimfi fimflmm. 38588.. «E «E 3: «E E: E: Esgfimzow a: s: ammfiéw 23.8w «$33 «5%on .832 mm 3 mm mm a a. 35 #83 Dawn: “fiat.” £93: nmmfiomb nmemK 98.1%: ammflmfi flaunts 3.30:. 28588“ «E s: «a «E s: 65:35 £385 £380 €3.30 Easeswbw £39K «9.38 «View 332% «93% mwvflfi 23$? #30? “£3? 325 t. a. a. 3 mm R mm mm 8 as...” “#83 €553 u 52%... .80. V 8 moocopowzu Hgomawwm 8865 5:38 a 3 £032 380th .80. N 8 cocouotmn 28$:me o: 8.885 5:38 a E 9532 088 one .33 m .aoxnh mam: 8chme 88:2: noumgov 935% 05 6% 96:3 * £30: @9555 £3.35 fimflofi Dana: swig 893$ 839$ @943. 98588“ DSVHE «SHE gnaw: 8T3? «sung 83% «SHE «Raw: 33am? §N§§¥ £3: £3: “SHE “mafia: «SHE 83m: 98$: nmmflp 839R 3.9:: a R mm a 8 mm 8 cm 2 9.5 M83 2 “53$: «83? “~35 gram: «8.3? «333. 8385 awash SHE owogeoéow n 53:63 “Rae? «$3 8ng 83$ «Rf? £33 835 £413 §§§§ 923m: fiflfi smflfi 83E £3? £335 £33 3385 £33. 323 m: S S 2 E 2 a : 2 ES x83 53% £525 £33 #:HNQ £2“me £3 Ham ASHE nwflowm firlém 283038 «Ram? #35 £320 823% $2.3? 293% 5%“: 2335mm «SHOE §N§§tw smash gimme “SHE «8 Han «mSHmmm $238 331% «8.33 RENE 82$ a w b o w v m N H n35 M83 .308 338%: 880:6 8% so 3 $83 H Mo Em H 886 .85 swag $8 583 .E 5223 .Amo. V 3 mooaeoucmc Hgomawu 3.865 5:28 a E mHBHB 380:5 .Amo. N 8 megawatt 38¢?me o: 8865 5:28 a 5 $032 2:8 2F .58 38wa wfims 8:85:me 8865 :oumEuw Educfim 05 8% £033 * 123 aim: 3380 033% 83m: 23mm» 2%?» Baggage“ £ng 533 32%”? pig £3ma 83; §N§§EV «Rama 33mg «83% £335 33% wwvfloww ESE am S mm mm R S Esau M83 83E aflofi nigh £396 ommflmo 513% ammflfi £33. 3385 Baggage“ 232m «333 comma €385 2&3? $385 9333 §N§§bw £35 #38 Swag 838m «SHE swag: 8%ku 3%: S 3 an R mm mm 8 $55 M83 $02583 E xenon? 124 Ame. V 8 muoceotww “:85:me 8865 5:38 a E ficfiofi 880:5 .Amo. N 8 cocaotfi €85:me o: 8865 5:28 a E 3032 08$ 23. .68 m .xoMPH mam: oogomuau 8865 :3ng 285: 2: Sta ago: * .m:.Hom. mEHNm. a3.Hmm. wifimm. «Erflmm. «Shawn. «Brfimm. m3.Hom. nmfiumcm. ouuohfiocuom wufifibm. «34.3. wofiflmfi. mofiflmw. wofiflhv. «mofiwv. mofifiom. «aoHOm. nacfifimm. Eitofimbv. «mofiumv. wooHQV. woova. amQHov. woodmhv. «werfiwv. accidom. wwo.H~m. waHcm. .8qu R 8 mm a mm 8 a om a mean M83 nmfifiom. afificm. afifiom. 2:.Hom. page? DNH.HOV. nofiHHv. nofiHmv. nofiva. ouoohfiofiuoa «29%. 83mm. «2.3m. @233 233%. £343 «:30 £2.30 fieAwm. §Es§¥ «8.th again. “8.3m. «8.3m. «mega. «89%. «woAS. «2.30. 3418. L83: wfi 5H 0H m8 V8 mH NH H8 OH “wean M83 nofiflmv. nofifimv. nofiHmv. nofifiov. nSHfix 0:.Hfim. ogrfiwm. an.H$. pangfima. Boozfiofiuow goofiflom. woofflam. whofimo. «woflmw. maoHvo. nmoHoc. adeqfloh. Quechua. anHmm. $333:me «Shag. .89qu acqfiov. «moHNh. amofiob. mfiofiflw. wm0...|.mw. «~9va. *woo..|...©w. x832 a w b o m w m N 8 “mafia M83 .308 330.8%: 388% 8:: :0 wow 33% .0 .«o Em H £828 88 Etta >38? .2 x6585 125 «REE EEEERFQ mo 58v 8 26 2%:56 8: «68893 «E 2885 :828 a 5 fiat»: 380th A3. N 3 855%“. “58¢?me o: 3865 5:28 a 5 «$32 08mm 2: .AoEmmgw .Amo. V 3 «8:80th “gomefi 203 88 Mo 38 93 «35 E 38.“ «.8255 Ho :8 Page. wfims 8chme 8885 shiatsu 3.85% 05 8% «.833 ... «8%2. «3A2. «:12. «8H2. «8H2. «23.8. «8A2. «3A2. «moxlé. sageoég «E «E «E «E E: «E «E «E «E :SEoEth «SHE. «2A2. «245. «292. «2A2. «3H8. «:98. «9H8. «3H8. «8:: «m mm Q E on 9. «v S 9. «wsi «83 «8A2. «3H2. «8A5. «8A2. «3H2. «8H2. «3H2. «89cm. «SQHON. 3838? «E «E E: S: «E «E «E «E n3 . H ac. EzEcwaV «2H8. «2.3m. «3H8. «Mafia. «8H3 «858. «£32. «3H3. «8.3m. «8:: a. 3. Q a. 2. 8 mm mm R «can «83 fiona. n«wo.flm. «8H3 £93m. «.8.H-. «23%. «8A3. «898. «238. 385558 «8A3 «89:. «892. «SHE. «8.45. «2H8. «31+.R. «2H3 «2.3m. §N§§tw «8.3m. «83.0 «8.3. «23.3 «8H2. «8.3m. «8.30 «8.3m. «3A3. «83: mm mm «m mm mm 8 8 mm «a fig «83 «82583 Z «689? 126 «BEEN Ecuatomxo mo 586 9 26 0:356 8: 3:88qu 8: .Go. V 3 moonouotmv Edouawu 88:5“ 5:38 a 5 «H892 Bogota Ame. N. 8 885:6 “52.33% on 8.82:: 5:28 a E «8&2 083 23. $32.56 0.83 8% 3 30m 23 Eco E 3.8-“ 3:825 no 58 38st wfimn 858$ng 386$ :ouwgov Buggy. 05 Bad «5:3 * 398. 89.2. 89:. SHS. 8A2. «.«oxlrz. Boomeoéa 3: ME .3: NE NE 3: Sawtcgwguvx «a «a «a «E «E «:38. «8:: 8 «m «m E mm mm «Sm «83 .602683 E «gunman? .Amo. v 8 moocouotwu EdamEmwm 98ch 5:28 a E «.882 Bahama .Amo. N & 8:809:38 “58$:me o: 3865 5:28 a E «.882 25% 2E. .88 «23:89 mam: 8ngme 8865 E5858 3.65% 2: Sta «5:3 * 127 «89:. «She. 888:. «89:. «8A2. 888:. 888:. 88...“? 888:. 985088 «8.3m. «8.8m. «8.8m. «8.3m. «8.8m. «8.3m. «8H8. «8.8m. «8.8m. «3.882% «3H3. «2H3. «NEE. «:9? «88S. «28%. «:88. «:88. «241%. «8:: R «a 8 «m «a «N a om a 88m «:83 889:. 83:. 8:.Jur:. 8:.H:. 88H? 8:.H:. «28:. 8:.H:. 83:. 285088 «8.44.8. «8.8m. «88%. «88.8. «8.Ho«. «889.. «8.8% «889.. «893. «3858‘ «2H8. «SHE. 83:. «28%. 873m. «28%. «2A8. «28%. «9.3m. 88: M: : 8 2 E 2 S : 2 Es: «.83 83:. 83:. 83:. 888:. 88.8“. 8883. 88.45.. «:88. 8:138. 2858? «8H3. «8A3. «883.. «89.3. «89.9.. «8H3. «2.3. «8.38. 8.“: §§§bw «2H8. «Quin. «9.3m. «EAR. «8.8m. «3.41:8. «SHE. *«8Ha. 898. .88: a w b o m « m N : 88m «:83 .288 33838 628.38 8:: no to: .8868 H :o Em H 88:: 88 E333 >283 .> 596%? 00 2 11 .Amo. V 3 80:88.2: “58:58.: 3885 5:38 a 5 80:2 580th .Go. N 8 888%: “58:58,: 0: 8885 5:58 a 5 888— 058 25. .88 «385:. 8:8 858:5? 8885 8:85: 28:88 05 8% 80:3 ... 8895. 8:98: 8:9:. 879:. 8:9:. 8:9:. 8:9:. 8:9:. 8:92. «88888 9889:. 889:. 8898. 8898. 8898. 8898. 8898. 8898. «898. «3885: «898. «898. 8:98. 8:98.4«898. «:98. «:98. «:98. «:98. 88: «m «m 8 :m om 8 8 t. 8 88m :83 8:92. 8:92. 889:. 889:. 9.89:. 889:. 889:. 889:. 889:. «888888 8898. «898. «898. «898. «898. «898. «898. «898. «898. «3.888: «:98. «:98. «:98. «:98. «:98. «:98. «:98. «:98. 8:98. 38: 8 3. 8 8 z. 8 8 8 R 8.8: :83 889:. 889:. 889:. 889:. 889:. 889:. 889:. 889:. 889:. 388088 «898. «898. «898. «898. «898. «898. «898. «898. «898. «3398:: 8:98. 8:98. 8:98. 8:98. 8:98. 8:93.. 8:98. «898. «898. 38: 8 mm «m 8 mm _m 8 8 8 88m :83 88883 > 8:55? 129 .30. V 3 88:20:88 880$:me 8865 5:28 a 5 8.882 858th A3. N 8 858me “58¢?me o: 8885 5:28 a E 8082 089. BE. .88 m 18:3. wfims 8:85:me 88:85 838328 888598 05 588 $083 ._. nmo.Hoo. .88....”00. avo.Hho. «vo.fibo. SQHQO. noono. §O.HS. nugfiofi. “EHHS. 3005:0588 wmo.Hbo. «aqfiao. wmQHoH. 83.48:. avonfi. namoHQ. swoo.H£. ESQHE. parqfibfi. Eiaofiwbw wmQHbo. «aqfiz. wwodvmfi. unqfiofi. wofifiom. «88%. wofifivm. wofiflom. ~8.me. .8qu 8 8 8 8 8m 8 «m 8 mm 88”: «:83 8:883 > 888% Ame. v 3 mooaeotmu :83?me 0832: 5:28 8 5 £052 880th A8. N & ooaohotmu 9:80:83 o: 8.888 8828 8 5 £032 088 2F .38 m .xoxsh mam: oofiomiwu 3885 8:833 @8983 05 ~88. 98:3 * 130 6 “we no H8 8 H8 6 H8 938 938 8H3 03.8 03% osgaoég 8%: as“: not“: new”: ha“? “$8 $8 in» $8 §§§bw 8382 8 Hz: 8 H8 8 H8 8H; «38 8H8 8th «38 82$ am 8 R 8 mm 8 8 S 8 85m 883 03% 31A? 3H: 3%.? 8H? 038 938 8H8 fl“: 983%de fine an“? 88.8 DES is 938 8;:er 8H8 8H2 §N§s§< $8.8 8H? 8H8 «was 8va 8H3 2H8 238 «398 3%: on 2 M: S 2 2 E a S 82m #83 nine 2? 23 2.11.3 84ch 8.33 £5.38 “no“? «3.8%. 285888 8“: nine a“: 833 92.3; 8.3186 «3“? 8.338 «9135 5:83:28 «38 8H: 8H: «23.2 «3&8 Sig «:38 8434 “188.88. Sui : 2 a w h o m v m 8.55 x83 .308 85828 Enact? ooh: :0 8% 33828 .0 Ho am H 8825 :28:on wwo 038383 >28?» .5 iguana“? 88:88 8080808098 mo 5808 9 020 03288 8: 38000308 8 \8 .Go. V 8 80080008820 E80E8wmm 08065 8828 8 E 0.80:2 08080885 Ame. N & 00:80th 0:82.:qu c: 08065 5:28 8 E 20002 0880 EH .A0E8=8>8 0803 888 go 300 93 3:0 .5 :03 8.803: Ho 0000 0.23038 mam: 00:89.28me 08068 828308 889808 05 0008 0000004 * anS cm HR: aw H03 aw Hoe ow H2: Db Hmofi aw H2: aw Ham aw me 200%80588 82 8? 8? 8). 8B 8B 8? 8? 8E Ezssfituw «E 82 H82 88 H9: 82 H9: 8: Hm: «a HE 82 H8: 83 H8: 2 8 H E 80:2 mm mm 8 mm 8 8 on Q 3 0858 883 aw Hem ab Hem ab Hmm ab H3 n6 Hma AK. H3 aw Hum AF Hmw ab Hmw 000938588 m 8? 8? 8E 88 8B 82 8E 82 8E Ezwzesmzuw wzHo: 8:32 8302 «EH? aeHofl «SHE 8EH02 83H03 5H8: 80:: S 3 Q 3 Q Q 8 08 am 0858 883 AK H8 AK H88 0» H; as HR 0» HR as HE fl HQ 08 HE 0» H80 9850588 88 .8 HQ 80 HQ QSHK 8H: 8H: 8H: 02H: 02H: 53385208 wflHmfl 5H5 «SHQS HSHVNH EH02 82sz 82H0: 8SHm: 82H? 80:: 8 S 08 mm 8 mm mm 8 cm 0858 883 .ABaécoov S 52083 132 88888 3808598 Ho £808 8 030 03288 Ho: 38080508 8 E 8H3 8H0: 8H8 meS 98888 88 8 E 8E 8E Eincfiwbw 8B 8? 88 8E 88:: 8 mm 88 R 0858 883 €380 5 £55 133 Ame. V E 880000.50 88¢?me 88:05 5:28 a 5 20302 3000me A8. N. 3 08000006 0:032me 0: 00865 5:28 a 5 20:2 0800 05. .602 m .523. mam: 0o§omea 008:2: «850300 figgm 05 00¢.» 30:0..— * «2252 «2?: «203: «2.162 «2.32 «:32 «24$: «3.3: «3082 BBEOEQ «:Ho: «232 «2.32 «232 «2.32 «01$: «22:: «2.22 «2082 §§§8w «8.112 «8052 «8.32 «2003 «2.3: «232 «222 «2.12.2 «23: «8:2 2 «N R 8 2 «N «N «a a 205.0 «283 «3me «2H: «2wa «2.08 «SHE «2H8 «2.03 «SH? «51...: 385808 «23: «2H2: «2H3 «3.03 «230 «2.1.? «SHE «22% «27:5. §.§«§«w «2.122 «2.32 «:32 «20.8 «20% «242% «2.1.2. «2H% «2.1% «8:: cm 2 2 2 2 2 3 2 2 mean «283 .5 H8 8E2 £02 63 «in «3ng «2:35 «033 o 985258 «82va «3 “8 «£08 £32 «8i: «3%? «mafia «gum; 2.0.2. §N=§§< «« HS. «« Hem «3% «3.2 «32 «2H? «3.1.? *«mffi: 9.0%. «8:: z 2 a w n o . w v m 2055 «283 .2008 39838 0:000:30 00.25 so 00% 33% H 00 Em w... 2825 «2285089 30 033256 «€200? .=> fic=0mm< 134 .Go. V 8 00020000020 2:85:me 0:00:05 5:28 0 5 2022 E00805 Ame. N 3 00:800wa 0:02.02:me o: 3868. 5:38 a 5 2022 08.8 0.3. .20: 3823. mam: 00:00:2me 0:00:22 E23300 2.20520 05 0020 2020A * «w 0me «N. 02N «0 0.2 «N. 0.80 «a 02.0 «a 03.0 «2022 «202.2 «a 0:2 385808 «202.5 «20% «20:8 «80:08 «8008 «8008 «80am «808m «8002 2322.202 «2008 «20.8 «205 «20c: «2008 «2008 «2008 «208m «20me .832 on mm «m mm mm E on 2. 2. :25”: «:83 «2008. «a 02m «0 02m «2020 «202:0 «:0mom «220.com «:08: «:02 98260508 «8002 «8020 «80:02 «2022 «80:2 «2002 «80mm «8008 «2020 assessow «2020 «20mm «2022 «20.22 «202.0 «2002 «8022 «20.08 «$0.8 82: t. 2. 2. 3. 2. 2. 2. 9. 2.. :05": «:83 «208: «202: «204.2 «200: «2022 «30m: «:02: «E08: «:02 283.0508 «202a «202m «20:2 «20:8 «202 «20.92 «2022 «3092 «20m: 33:95.2. «2020 «20.2 «2022 «80.8w «S002 «80:02 «202 «80E «80o: 32: 2 a on 2 «m mm mm :m cm 25": «:83 .Agscuaoov E 230%? 135 .89 V 8 mooceotmu «58$:me 88:2: 5:38 a 5 «5:2 880th .30. N 8 ooaouotmu “=85:me 0: 88:2: 5:28 a 5 20:2 2:8 2; .38 «ahead. mafia: 853$:me 8865 «83336 9.85% 05 Hoax «.833 * mm Howm ab H 5m an HEN 3 H mum we H mom an H vow «a Hoom Saga—0:38 womHowm momH aha momH mum ammH EN «~me EN mva 3N wva mum EBEQEEQV m3 HNwN «a H SN «2 H 5N «2 Ham «3 Howm «3 H arm «3 HEN 39:: «« 8 6 8 «m «m 3 «Sn «83 «82583 => 52%? 136 .Amo. V 8 000202028 28022:me 000065 5:28 a 5 20202 E02025 .Go. N 8 002020220 “Swocawmm 0: 0200222 52300 a 5 20:02 08$. 05. .202 362:2. 2222 0020022me 020065 «820300 22282 05 2020 2020A ._. «0002 «03.2 «03.2 «03.2 «0002 «0002 «0002 «0002 «0002 285.228 «03.8 «03.2 «0002 «03.2 «03.2 «03.2 «03.2 «0002 «0002 «52322. «03.8 «0002 «0002 «03.2 «03.2 «0002 «0002 «0002 «0002 «8:2 2 3 mm «m 2.. mm a om 2 2250 283 «03.2 «03.: «0002 «000a «0000 «003 «0000 2.000 2003 985808 «03.2 «0002 «000: «03.2 «000m «000m «0000 £0000 20000 2:28.22 «0002 «0002 «03.2 «0002 «000a «0000 «003 «0000 «0000 «8:2 2 2 2 2 2 2 2 2 2 25.2 .283 «003. «030 2033 8.002 «2.30 «2.02.. «2.30 «2.02. «3.08. 2828808 2000.. £030 ««000« £033 «2.08.2 «2.30 «2030 «2.30 «8.08. «5.532% «030 «003. «000m «0003 3002.3 «2.02.2 «2.02. «3.02.. *«202. .522 a w 0 0 m « m N 2 2250 283 .258 358028 22020.28 002.: so 00.2 2.2% H «6 Em H 280:8 20202005 02250 0.625830 >200? 22> 20:023.. .Amc. V & 00802028 8822220 008:2: SE28 0 E 20202 E02025 .Amo. N 3 005.2026 2802222 0: 0:00:02 5:38 a S 20:02 08% 0E. .20“ 0.2023. 2%: 00:02.2:me 028222 2820300 220520 05 2020 2023 * wmAHoAv mméfimdv wmAHodm mvéflodm wméfiodm «HAHN? 2.3.8.9” wméflbwm 3.2.06.3 0200>Eo§ow 00. Hwév 00. H03. mm. dev 0w. ”2.9. mm. Hmdm an. ”.0me no. Noam an. Hfibm an. “0.0m 233020.202 3. H23 2.. HQ? mm. thm 0w. Hodm aw. Hm.wm mm. HbSm «a. dem moéflodm 3.3..N6m .803: E mm mm a om 2. 2. 2. 2. 2252 283 7 Séfimim «NAHmdm an.~Hm.mm noéfioém moAdem moAHodm Eéflcdm ww.Hw.wN «QHmfim 0000>Eofifim B 3. Hm.mm no. Hwém no. Hfivm «WHNJH mm. dew mm. H5; 3. dem «WHQQN «QHCdN 23.305032 woAHvém aw. Hbdm 0m. Hodm nao.Hm.Nm «OAHmém «a. Hodm woéfiadm 0QH2¢N «onyxwm .8032 2. .3. 2. 2. z. 2. mm 2 R 2052 .283 «WHfifim «QHmém 34.5mm 5.0.5.3 «QHNéN wwdem «WHmdN 3. No.3 .3. Hodm 02002802200 ~©.Hfi.wm «WHmfim 00%de £0.92 «QHoém «QHWmN ww.Hm.~m we. Hfimm 0w. HmAN 23.592202 magnmfim am.Ho.om «afflwfiw 23%qu 00.03.23 «QHmdm «QHQNN moAHwAN 8.3.6.3 28:2 0m 2 «m «m 2 2m 8 2 mm 2252 283 282583 22> 22229.. 138 .Go. v 8 888058 “58$:me 882:: 5:28 a 5 $052 Bogota .Amo. N 3 856:6 258.3:me o: 9885 5:38 a 5 $032 08am 2? .33 m ROME. mam: 898$:me 8865 dosages 32:83 2: Sta 30:3 * avAHmSv mméfimfiv mvéflvdv wméfiofiv 8.36.3 mwéflwdv wifififimv wméfimdv wmAH©Av Boo>Eocaow wméfimfiv «méfibdv maéfifiov moéfivdv aw. Hwév aw. Hfivq 8. H5? ab. No.3 3. Hvdv :SEQEEQV aw. Hwéw ab. Hmdv an. Hmfiw .3. Hfifiv an. HN.3 ab. H59“ 3. Hodv ah. Hmdv an. H5; .893: 8 8 s 8 am mm R on R asa x83 .ABsécoov :5 xenon? HICHIGQN STQTE UNIV L BRQ i leif MINIMUM HM! rl if! W l“ 3 1293010259004 v, \ .