THE EFFECT ON THE CEREAL my 7 . 7 same 0F PLANTING on; e . . ~ WITH A- COMPANION CROP; [ . Thesis for the Degree. of M. S. , MlCHlGAN SYATE UNWERSiTY ' ALANLSAWYER ~ ' 1976 - ' ....... ...... ..... .' N‘ I, I: CV} LA: ABSTRACT Il * THE EFFECT ON THE CEREAL LEAF BEETLE OF PLANTING OATs WITH A COMPANION CROP by Alan J. Sawyer A review of the literature concerning the many forms of companion planting is presented, and a unifying definition is proposed. Results of a two-year study of the effects of interplanting spring oats with companion crops of alfalfa (for hay in succeeding years) and spring-planted (and thus, non-vernalized) winter wheat (for a trap-crop) are reported. Activities were aimed a quantitatively measuring oat growth and yield and evaluating cereal leaf beetle (CLB) densities, mortality factors, and damage to the crop. The experimental treatments affected soil temperature, oat stem density and height, foliage wet weight per ft2 and per stem, leaf sur- face area, and grain yield; but these effects were extremely variable from field to field and both within and between years. Plant dry matter accumulation and nitrogen content, and soil moisture were essentially unaffected by the treatments. Few differences attributable to companion crop effects were found in CLB densities, age specific mortalities, parasitism of CLB eggs and larvae, or feeding damage. Emergence of the second generation of Tetrastichus julis, a larval parasite of the CLB, occurred earlier in the companion-planted fields but this is unlikely to significantly improve its effectiveness as a biological control agent. Alan J. Sawyer Alfalfa established in the spring with an oat companion produced the same amount of forage at first harvest as alfalfa established in the fall without oats, but the species composition of associated weeds was quite different. Alfalfa weevil densities were too low to evaluate the effect of establishment methods on this pest. Quantitative models of oat growth and CLB population dynamics should be, for the most part, equally valid whether oats are grown with an alfalfa companion crop or in pure culture, two methods commonly used in Michigan agriculture. THE EFFECT ON THE CEREAL LEAF BEETLE 0F PLANTING OATS WITH A COMPANION CROP by Alan J. Sawyer A Thesis Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Master of Science Department of Entomology l976 ACKNOWLEDGEMENTS I wish to express my sincere appreciation to Dr. Dean L. Haynes for his guidance and support while serving as my major professor, to Drs. George N. Bird, Peter G. Murphy, Richard J. Sauer and James A. Webster for serving on my guidance committee, and to Marcia James Sawyer for her constant encouragement. I also thank the National Science Foundation's Graduate Fellowship Program for financial support. ii TABLE OF CONTENTS Page LIST OF TABLES ........................ iv LIST OF FIGURES ........................ ix INTRODUCTION . . . . . . . . . . . . . . . I ......... 1 LITERATURE REVIEW ....................... 2 Companion Planting .................... 2 Effects of Legumes on Other Crops ............ 5 Other Interactions of Companion Crops .......... 5 Crop Heterogeneity and Pest Damage ............ 6 METHODS ............................ ll Crop Component ...................... ll l974 ........................ ll l975 ........................ l4 Insect Component ..................... 2l l974 ........................ 21 l975 ........................ 22 RESULTS AND DISCUSSION .................... 24 Oat Growth Under Companion Cropping ........... 24 Soil Moisture ...................... 55 Soil Temperature ..................... 57 Plant Nitrogen ...................... 70 Insect Response to Companion Cropping .......... 8l CLB density ..................... 81 Parasitism of CLB larvae .............. 95 Parasitism of CLB eggs ............... lOl CLB survival .................... l04 Field cage study .................. l09 Emergence of I, Julis ............... ll2 Feeding damage ................... ll5 Alfalfa in the Second Year ................ ll9 SUMMARY AND CONCLUSIONS .................... 125 LITERATURE CITED ....................... l29 APPENDIX A - Gull Lake Degree-day Accumulations ........ l36 APPENDIX B - Recovery Efficiency of CLB Pupal Cells ...... l4l iii Table l. Acreage, yield, and market value of pure and mixed grains grown in Ontario in 1971 (data from Ontario 1971) ....... 2. Relationship between variances (52) and means (x) for oat growth parameters ....................... 3. Growth parameters (Y': SE, N = 30) for oats grown in pure culture in field 9-16, 1974 .................. 4. Growth parameters (Y': SE, N = 30) for oats grown with alfalfa in field 9-17, 1974 .................. 5. Growth parameters (Y'i SE, N = 30) for oats grown with non-vernalized wheat in field 9-11, 1974 ........... 6. ANOVA results for 10910 transformed oat stem density, plant wet weight and height, and per-stem weight for 13 sampling days in l974 ......................... 7. Oat growth parameters (Y': SE, N = 30) for all treatments at peak larval time and final date of observation, 1974 and 1975 ............................. 8. ANOVA results for oat growth parameters (10910 transformed) at the time of ;eak larval densities and at the end of the season in 1974 and 1975 .................... 9. Comparison of average growth responses for the three crops at the time of peak CLB larval densities and at the end of the season in 1974 and 1975 .................. 10. Length and calculated surface area on July 4, 1974 of the upper three leaves of oat plants in the three treatments . . . 11. Length and calculated surface area on June 30, 1975 of the upper three leaves of oat plants in the three treatments . . . 12. ANOVA results for squared leaf lengths (logio transforma- tion) of the uppermost three leaves in 1975 using the method of unweighted means .................. 13. Mean squared leaf lengths (10910 transformation) of the LIST OF TABLES uppermost three leaves in the three treatments in 1975. Estimated variance = .001 with 687 df ............. iv Page 26 3O 31 32 33 4O 41 42 47 48 49 51 Table Page 14. Oat head density (#/ft2), air-dry head weight (g) and estimated_grain yield (bu/a) for the three treatments in 1974 (X t SE) ....................... 52 15. Oat yield for the three treatments in several experi- mental fields in l975 .................... 54 16. Soil water content (% of dry soil weight) in the top inch of soil for the three crops over a five-week period in l975 ........................ 56 17. Soil and above-soil temperatures (0C) recorded at Gull Lake, 1975 ....................... 59 18. ANOVA results for soil temperature at Gull Lake in l975 ............................. 60 19. Soil temperature (0C) for the three treatments in each field at Gull Lake, averaged over depths, row placement and dates (n = 12, S = 3.27 with 74 df) ...... 61 20. Accumulation of total above-ground dry matter (g/2 ftz) at Gull Lake plots, 1975 (includes oats, alfalfa, broadleaf weeds and grassy weeds) .............. 65 21. Accumulation of above-ground oat dry matter (g/2 ft2) at Gull Lake, 1975 ...................... 66 22. Accumulation of above-ground grass dry matter (g/2 ftz) at Gull Lake, 1975 ...................... 67 23. Accumulgtion of above-ground broadleaf weed dry matter (9/2 ft ) at GOIT Lake, 1975 ................. 68 24. ANOVA results (LogTo transformation) for dry matter components in the Gull Lake plots, 1975. Entries are mean squares and significance level ........... 69 25. Nitrogen content by date (g/2 ftz) for all plant components (0 = oats, G = grasses, BLH = broadleaf weeds, A = alfalfa) in the Gull Lake plots, l975 ....... 72 26. Total above-ground plant nitrogen (g/2 ft2) at Gull Lake in l975 ......................... 74 27. ANOVA results for above-ground plant nitrogen compo- nents. Entries are mean squares, (degrees of freedom) and significance level .................... 75 Table Page 28. Percent nitrogen (by weight) of oat heads on August 4, 1975 at Gull Lake. Samples were two linear feet of grain ..... 8O 29. Density (no/ftz) of CLB life stages in the pure oats plot of field 9-16 in 1974. (X i SE, n = 30) ........... 83 30. Density (no/ftz) of CLB life stages in oats planted with alfalfa in field 9-17 in 1974. (X i SE, n = 30) ....... 84 31. Density (no/ftz) of CLB life stages in oats planted with wheat in field 9-11 in 1974. (X t SE, n = 30) ........ 85 32. Comparison of the 1974 peak CLB egg and larval densities [log10 (no/ft + 1)] in the experimental fields ........ 86 33. Densities (x': 5.0.) of CLB pupal cells and emerging summer adults in l974 ........................ 88 34. Density (no/ftz) of CLB life stages at Gull Lake on June 13, 1975 (x e SE, N = 30) ................... 89 35. Comparison of CLB egg and total larval densities (/y + .5 transformation) in field 9-10 on June 13, 1975 ........ 9O 36. Larval densities (no/ftz) estimated from sweepnet catches and estimated seasonal population for each plot in 1975. See text for methods of calculation .............. 92 37. CLB pupal cell and emerging adult densities at Gull Lake in 1975. (x i 5.0.) ..................... 93 38. Mean egg and larval densities on June 10, 1975 and mean pupal densities on July 29 at Galien, and error mean squares from analyses of variance ............... 94 39. Percent parasitism by I, julis of large CLB larvae collected for dissection in 1974. Numbers in parentheses are the number of larvae dissected .................. 96 40. Percent parasitism of CLB larvae by I, julis in 1974 as measured by examining pupal cells collected in soil samples. (15 l/2-yd samples per field) ................ 98 41. Estimates of larval parasitism at Gull Lake in 1975 based on sweepnet collections and soil samples. Numbers in paren- theses are total larvae dissected to arrive at each estimate . 99 vi Table Page 42. Percent parasitism of CLB larvae by T. julis at Galien in 1975 as measured by examining pupal cells collected in soil samples( (8 l/2-yd samples per plot) ......... 102 43. Percent parasitism of CLB eggs by Anaphes flavipes in the three crops at Gull Lake in 1974. Numbers in parentheses are the number of "viable eggs" collected in 30 ft2 of foliage (see text) ...................... 103 44. Total production (no/ftz) of CLB eggs, larvae, pupae, and summer adults in the Gull Lake experimental plots in 1974 and 1975 ........................... 105 45. Survival of CLB life stages at Gull Lake. . . ........ 108 46. Survival of CLBs from egg to pupa in the field-cage study at Gull Lake in l975 ..................... 111 47. Cumulative percent emergence of I, julis adults in ten 1-yd emergence traps placed in each experimental oat field in l974 ........................ 114 48. Emergence (number and cumulative percent) of second gener- ation I, julis adults in 30 l-yd emergence traps placed in each crop in l975 ...................... 117 49a. ANOVA results (mean squares) for damage to the tOp three leaves of oats in 1974 and 1975. Error mean squares are followed by their degrees of freedom (in parentheses) . . . . 118 49b. Seasonal larval density and subsequent feeding damage to oat leaves in the experimental fields in 1974 and 1975 [Mean number of CLB larval feeding scars per leaf (both years) and estimated percent of leaf surface area which was green (1974 only), averaged over the top three leaves] ....... 120 50. 1975 measures of plant growth in the alfalfa fields estab- lished at Gull Lake the previous year by two methods (x i S.E. n = 10) ......................... 121 51. Total wet weight (g) of aerial portions of weeds in ten l-ft samples in the two alfalfa fields on May 29, 1975 . . . 123 2 52. Number and size of alfalfa weevil egg masses in ten l-ft foliage samples taken in the two alfalfa fields on several dates in 1975. Each entry gives the number of eggs in an eg mass; masses enclosed by parentheses are from the same ft sample .......................... 124 Table Page Al. Degree-day accumulations at Gull Lake for 1974 ........ 137 A2. Degree-day accumulations at Gull Lake for 1975 ........ 139 81. Recovery of CLB pupal cells ................. 143 B2. Analysis of variance ..................... 145 viii Figure 10. 11. LIST OF FIGURES Gull Lake crop map for 1974 ............... Gull Lake crop map for 1975 ............... Experimental plots at Collins Rd. (A) and Galien (B) in 1975 ........................ The relationship between means and variances for oat wet weight (g/ft2) in 1974 .............. Parameters of plant growth in the pure oat field in 1974 . . ...................... Stem density of oats in the experimental plots in 1974. Each point is the median of 30 observations. (Vertical axis log scale) ............... Plant height of oats in the experimental plots in 1974. Each point is the median of 30 observations. (Vertical axis log scale) ............... Net weight of oats in the experimental plots in 1974. Each point is the median of 30 observations. (Vertical axis log scale) ............... Net weight per stem of oats in the experimental plots in 1974. Each point is the median (plus 1) of 30 observations. (Vertical axis log scale) ........ Total precipitation for April-July and maximum snow depth during March and April at Gull Lake for the two study years. Climatological data from U.S. Dept. Commerce, National Oceanic and Atmospheric Adminis- tration ........................ Soil moisture trends in the top inch of three fields at Gull Lake in 1975. Each point is an average over the three crop types ............... ix Page 12 15 17 29 35 37 39 45 58 Figure Page 12. Dry matter accumulation on experimental plots at Gull Lake in 1975. Each point is an average over nine plots (except broadleaf weeds, where each point is an average over three fields) ......... 64 13. Total plant-sontained nitrogen (mg/2 ft2) in each plot over D>42. Each point is a single sample; lines were fitted by least-squares ........... 71 14. Percentages of total above-ground plant nitrogen per 2 ft2 in each plant component, by date. Each point is a mean of nine observations (except for alfalfa, where each point is a mean of three observations) ..................... 78 15. Percent nitrogen (by dry weight) of total above- ground oat biomass for each treatment on several days. Each point is a mean of three observations. Vertical bar is i 1 standard deviation ......... 79 16. The relationship between the variance and mean for CLB egg and larval densities per ft2 in oats at Gull Lake in 1974 (both axis log scale). Each point is a mean of 30 observations ........... 82 17. Percent survival to the pupal stage of eggs ovi- posited by adult CLB caged for one week in the three crops at Gull Lake in 1975 ............ 110 18. Cumulative percent emergence (probability scale) of I, Julis adults from the experimental plots in l974 .......................... 113 19. Occurrence of CLB larvae and cumulative percent emergence of second generation I, Julis in field 9-17 (0A) in l974 ................ 116 INTRODUCTION This research was undertaken to investigate what effects, if any, growing oats with a companion crop might have on the cereal leaf beetle (CLB), Oulema melanopus (L.)1, a potentially serious pest of spring seeded small grains. The term "companion crop“ will be defined below, and it will be seen as well that very little research has been done on the influence of companion cropping on insect populations. Descriptions of the cereal leaf beetle, its biology and its interaction with oats may be found elsewhere (Gage 1972, Haynes 1973, Jackman 1976). The companion crops used in this present research were alfalfa and spring-seeded winter wheat. A winter grain, if not exposed in the seed- ling stage to low temperatures (iLg,, is non-vernalized), will remain procumbent and will not produce grain (Salisbury and Ross 1969). The interseeded alfalfa was intended to be cut for hay in succeeding years after the harvest of oat grain the first year. The wheat was originally intended to serve as a trap crop to draw beetles away from the oats. Trap crops and this method of hay or forage crop establishment will both be discussed below as forms of companion planting. In 1961 Peters stated that the "decreased demand for oats for grain, combined with recent success in the herbicidal control of weeds in new legume seedings has raised the question of the advantage to be gained by continuing the practice of using an oat companion crOp." Hume gt_al. (1969), too, found the use of herbicides to be a feasible alternative to the use of an oat companion crop for alfalfa establishment. But now, 1Coleoptera: Chrysomelidae indications of increasing oat production in Michiganz, coupled with rising costs for nitrogen fertilizer and fuel and a growing concern over excessive pesticide usage in agriculture may indicate that it is again time to emphasize the values of companion planting. A review of the literature concerning companion planting is now presented to introduce the subject and demonstrate its scope and import- ance. LITERATURE REVIEW Companion Planting Denisen (1958) has defined companion crops as "two non-competitive crops grown in the same area at the same time. One is short term and the other of long duration." While he was referring to vegetable crops grown in the home garden, the same definition applies to the case of a forage crop established by interplanting the grass or legume with a small-grain companion, or "nurse", crop. Kilcher and Heinrichs (1960) state that this has been a long established agricultural practice. Numerous articles report the value of companion cropping in legume or grassland establishment in experiments conducted around the world (Dijkstra and DeVos 1972, Frame et_al, 1972, Haggar 1970, Kust 1968, Lawrence 1970, McGowan and Williams 1971, Olsen and Tiharuhondi 1972, Shah 1965, Zavitz 1927). These authors report such benefits as rapid 2Michigan oat acreage has increased an average of 5% per year since 1972. Data are in several publications of the Michigan Crop Reporting Service, Michigan Department of Agriculture (Michigan county estimates, Field Crops, 1959-1972; Michigan Agriculture Statistics, June 1974; County estimates-oats, April 1976). crop establishment, protection against wind and water erosion, lessened weed invasion, increased forage yield in the seeding year, improved harvestability, increased forage nutritional quality, and additional income from the harvested grain. Flanagan and Washko (1950) found that a good small grain nurse crop for legume establishment is one which allows for sufficient light penetration by having short stems and a low stem density; some of the newer oat varieties were highly rated. A few authors (Kilcher and Heinrichs 1960, Lawrence 1967, Peters 1961) report adverse effects of using companion crops, especially when soil moisture is limiting or when the perennial crop has a high cash value. Small grains are also interplanted for a mixed grain harvest. In 1971, 928,364 acres of mixed grain (oats and barley) were grown for feed in Ontario. The average yield of the mixed grain was higher than for either of the pure grains and the market value, both per bushel and per acre, was intermediate (Table 1) (Ontario 1971). Results of mixed grain experiments usually range from no effect to moderately increased grain yields, particularly with an oat/barley combination (Morrish 1934, Petrov 1968, Stanton 1929, Syme and Bremner 1969, Zavitz 1927). Other reported effects include increased resistance to lodging and disease (Petrov 1968) and higher straw yields (Zavitz 1927). Arny et a1, (1929) give several advantages of growing flax in mixture with small grains, including higher flax yields under favorable moisture conditions, in- creased bushel weights of flax seed, wheat and oats, improved weed control, easier flax harvest (less matting of stems), and a protection against total loss when wheat stem rust occurs since flax is unaffected. Table 1. Acreage, yield, and market value of pure and mixed grains grown in Ontario in 1971 (data from Ontario 1971). Grain Acreage Yield Market value (bu/a) ($/bU) ($/a) Oats 650,430 57.5 0.73 41.98 Barley 385,651 54.0 0.98 52.92 Mixed 928,364 61.0 0.85 51.85 (Oats and Barley) In experiments of planting cotton in strips with other crops (see also p. 8), Gouder and Patil (1971) found that although the yield of seed cotton was reduced by interplanting it with wheat, the monetary returns were higher due to the harvest of grain; and Robinson gt_§l: (1972c) found no difference in cotton yield or quality when it was grown in strips with corn, soybeans, alfalfa, peanuts, or sorghum. Effects of Legumes on Other Crops The beneficial effect of legumes on soil fertility has been realized for centuries, but the actual discovery of nitrogen fixation in root nodules was made in 1887 (Loehwing 1937). In 1914 it was known (Westgate and Oakley 1914) that bacteria were responsible for the fixation in nodules, and that other plants could make use of this added nitrogen. Research on the effects of leguminous companion crops on the yield and quality of other crops is of perennial interest and the literature is plentiful (Agboola and Fayemi 1970, Chestnut 1972, Christozov 1968, Evans 1916, Haggar 1970, Lyon and Bizzell 1913b, McClelland 1928, Petrakieva and Naidenov 1968, Son 1969, to name but a few). Cowling and Lockyer (1967) found that "pure grass swards re- quired more than 200 1b. of fertilizer nitrogen/acre/year in order to yield the same amount of nitrogen (in hay) as (an unfertilized) grass/white clover sward." Other Interactions of Companion Crops There are other root interactions which may determine the bene- ficial influence of one plant on another. Lyon and 3122611 (1913a) found that the growth of crop plants was often stimulated by later planting secondary, non-leguminous crops with them, and suggested that nitrogen is excreted by the roots of many young plants. Loehwing's review (1937) thoroughly covers the early literature on the subject of root interactions. Roots excrete many organic and inorganic sub- stances, including amino acids, B-vitamins, etc. which may provide nutrients for soil micro-organisms. These microbes may in turn affect the growth of plants by their own by-products, which include nitrates and anti-biotics (Anonymous 1960). There is also evidence from radio- active tracer experiments that plant metabolites such as amino acids and carbohydrates and inorganic phosphorus and sulphur excreted by one plant may be directly absorbed by other plants of the same or other species which grow nearby (Grodzinskij 1969). Thus, Grodzinskij says, "plants in close communities may have a common exchange pool of free organic and mineral substances via their roots." Crop_Heterggeneity and Pest Damage The relationship between ecosystem diversity and stability is an interesting yet unresolved ecological problem (see Pielou 1974 for a discussion). Ecological theory (Margalef 1968, Odum 1971) and some investigations (ELHJ Pimental 1961) indicate that there is a direct relationship. However, there is at least one entomological study which reports the reverse (Murdoch g__al, 1972). Rudd (1964) discusses modern agriculture's tendency to reduce the biotic complexity of the landscape. His examples of biotic simplifi- cation include the grain producing regions of North America where thousands of acres are sown in monoculture, and the large pine plant- ations of the southeastern United States. Other agricultural crops are commonly grown in monoculture in fields of 40 to 200 acres. The reason for this is economic. In such simplified ecosystems, higher productivity is achieved by maintaining the system in an early success- ional stage. Furthermore, most of the energy and mineral resources are funneled into the desired crop, resulting in higher yields (Odum 1971). Temporal diversity is also reduced in many cases by eliminating crop rotations. These practices expedite farming operations and require less extensive investments in machinery. While data on the effect of biotic simplification of croplands is neither complete nor wholly conclusive, the work of many researchers indicate that pest problems are intensified in monocultures (Beirne 1967, Franz 1961, Pimentel 1961, 1970, Rudd 1964). Beirne (1967) suggests reasons for increased pest problems in sim- plified environments: (1) Since most natural enemies of pests have requirements for reproduction and survival (such as alternate hosts and oviposition or overwintering sites) that exist other than where the pests are, many such enemies are eliminated or disrupted by mono- cultures. (2) Pure crop stands provide abundant food, which reduces competition for food and the mortality and energy expenditures usually associated with searching for food. Thus the "intrinsic rate of natural increase" of the pest is little inhibited. (3) Non-pests that are able to exploit the above conditions may become pests and minor pests may become major pests. Planting two or more crops in alternate strips or in adjacent fields has been investigated, with varying results, as a means of reducing insect damage. The intent is usually either to increase environmental heterogeneity to enhance biological control, or to use one crop as a trap to attract and hence remove pests from the other crop. Trap crops have been used successfully in preventing lygus bug damage to cotton (Sevacherian and Stern 1974, Stern 1969, Stride 1969). Alfalfa cutting has been manipulated to produce strips of different ages for preserving natural enemies of pests (Schlinger and Dietrick 1960) and for attract- ing and then destroying (with insecticides) the pests themselves (Scholl and Medler 1947). Some experiments with cotton strip-planted with other crops (DeLoach and Peters 1972, Robinson gt_ 1. 1972a, b) have produced inconclusive results which add to the uncertainty concerning the relation of plant diversity to pest numbers and population stability. Poston and Pedigo (1975) have shown that leafhoppers and plant bugs migrate from cut alfalfa to adjacent soybeans, and so under certain conditions it may be unwise to plant these crops in close proximity. In mixed crop plantings, too, results have been variable. Petrov (1968) reports that "barley and wheat suffer less from diseases when grown in mixtures than when grown alone," but Arny gt_al, (1929) found no effect on plant diseases such as flax wilt and stem rust of wheat when growing mixtures of flax and small grains. Atsatt and O'Dowd (l976) introduce the concept of plant defense "guilds", in which plant associates function to reduce herbivory. The authors cite several examples in which guild members serve as insectary plants (harboring herbivore parasites and predators), as repellant plants, and as attractant-decoy plants. The popular "organic gardening" literature makes frequent reference to companion planting. In this realm, two different meanings have been given to the phrase "companion planting", both dealing with the effect on pests: (l) the interplanting of aromatic herbs or flowers among crop plants to repel insects; and (2) the interplanting of crop plants which "like each other" or somehow offer mutual protection from insect pests, and segregation of crops which "don't like each other" or aggra- vate insect problems when grown together. There is usually no attempt in organic gardening literature (£49. Gillespie 1964, Johns 1966, Kappel 1973, Olds 1959, Philbrick and Gregg 1967, Philbrick and Philbrick 1963, Young 1973) to document these claims of reduced insect damage due to companion planting. Whenever controlled experiments are mentioned, literature citations are not given (Hunter 1971, J.C. 1973, Tyler 1970). This folksy situation has been aptly described by Hunter (1971): "Whether it is a matter of composition, exudate, or other influence as yet unexplored, there is an accumulation of practical experiences of gardeners concerning helpful and harmful combinations (of plants)" According to Beirne (1967), control measures reduce pest damage in one of two ways: either "by destroying the pests or by protecting the victim from attack." Ideally, victim-protecting controls which remove the possibility of damage should be preferred to pest-destroying methods which depend on regular reduction in pest numbers whenever damage is about to occur or already has occurred. However, in practice, pest- destroying controls are by far the most widely used. To bring together the various usages of the term, I propose that companion planting may be broadly defined as: 10 The purposeful inter-planting or juxtaposition of two or more crops with the aim of benefiting from such a practice. These benefits may include enhanced crop establishment and development, increased yields, soil improvement, reduced production costs, or reduced crop losses to pests. If companion planting leads to reduced pest damage, whether through an increase in environmental heterogeneity, in- creased plant vigor or by some other mechanism, then it may be con- sidered a "victim-protecting" control measure and as such may be desirable. It is clear that little research has been done to determine what effect specific companion crop combinations have on their associated pest complexes. To that problem this research is addressed. METHODS Research was done over a two-year period in the summers of 1974 and 1975, primarily at the experimental farm of Michigan State University's W.K. Kellogg Biological Station on Gull Lake at Hickory Corners (Kalamazoo Co.), Michigan. In 1975, additional plots were maintained at the Depart- ment of Entomology's East Lansing campus research area (hereafter called Collins Road) and on a cooperator's (K. Bohn) farm at Galien (Berrien County), Michigan. Activities were aimed at quantitatively measuring crop growth and yield and evaluating insect densities, mortality factors and damage to the crop. Crop Component 1224;. Field 9-16 (Fig. 1) at Gull Lake was sown on April 20 with 'Rodney' oats (2 bu/a). The adjacent fie1d, 9-17, was planted with a mixture of 2 bu of oats and 12 lbs of 'Vernal' alfalfa to the acre. Nearby, field 9-11 was planted with 2 bu of a 50:50 mixture of oats and 'Genesee' wheat. These treatments will be denoted by the abbreviations 0, 0A, and OH, respectively. The fields were about 2.5 acres in size, topographically flat, and with a Fox loam soil type (Perkins and Tyson 1926). The day before planting, 150 lb/a of 6-24-24 fertilizer was applied. Wheat and oats were premixed and drilled in 7 inch rows. In the OA field, the alfalfa was drilled separately. Ten study plots 50 ft X 100 ft were established in each of the three 11 12 huzoz a mh42 Stems/ft2 Height Wet weight/ft2 Weight/stem (N0.) (cm) (9) (9) 5/17 326 l8.3tl.l 8.1i0.2 2.4:0.2 0.13:.005 5/21 415 l6.7:l.1 10.1:0.3 4.2:0.3 0.25:.01 5/25 500 21.2:1.0 14.5:O.3 8.4:0.5 0.40:.02 5/29 569 29.3i1.9 18.6:0.4 20.3:l.3 0.69:.03 6/3 677 35.6:l.7 23.4:0.5 45.7:2.4 1.28:.06 6/7 797 31.5:2.0 31.9i1.0 67.3:4.3 2.20:.10 6/12 923 31.1:1.4 36.2:0.1 82.8:5.2 2.66:.10 6/17 1023 29.8:l.3 27.5:1.3 125.1:6.l 4.20:.15 6/21 1137 29.3:1.4 39.3i1.6 127.928.l 4.36i.26 6/25 1220 28.1:1.5 44.2:l.7 l40.0:7.2 4.98:.27 6/28 1296 26.5:1.3 48.1i1.2 142.4:6.8 5.38:.26 7/2 1433 23.7:l.2 61.3:l.7 139.4:7.4 5.87i.23 7/5 1543 23.0:1.3 66.2:l.5 121.4:6.7 5.29:.25 31 Table 4. Growth parameters (X t SE, N = 30) for oats grown with alfalfa in fie1d 9-17, 1974. Date OD>42 Stems/ft2 Height Wet weight/ft2 Weight/stem (N0.) (cm) (9) (9) 5/17 326 17.7i1.5 9.0:O.2 2.6:O.3 0.14:.006 5/21 415 22.0:1.4 10.5i0.2 5.0i0.5 O.22i.01 5/25 500 19.4:1.1 14.5i0.4 9.5:0.6 0.49:.02 5/29 569 31.1:1.3 20.3:O.4 24.6i1.3 0.79i.03 6/3 677 34.5i1.5 25.7:O.7 50.7:2.8 1.47:.03 6/7 797 33.2i1.5 30.7:O.8 67.4:4.4 2.03i.06 6/12 923 34.1:1.9 38.3:l.2 94.316.5 2.76:.12 6/17 1023 21.5i1.9 36.8il.6 87.8i4.7 4.08:.24 6/21 1137 24.8i1.3 43.611.5 125.8i7.7 5.07i.22 6/25 1220 23.9:1.3 50.2:2.1 147.2:8.3 6.16:.29 6/28 1296 23.1il.2 52.4:1.3 160.1i10.1 6.93:.21 7/2 1433 21.3:1.0 69.4:1.5 144.6:8.1 6.79i.31 7/5 1543 20.521.2 71.2i2.2 147.2:10.8 7.18:.32 32 Table 5. Growth parameters (X t SE, N = 30) for oats grown with non-vernalized wheat in field 9-11, 1974. Date °O>42 Stems/ft2 Height Wet weight/ft2 Weight/stem (N0.) (cm) (9) (9) 5/14 279 23.8i1.5 8.9:0.1 2.4i0.1 0.10£.OO4 5/17 326 26.2:1.2 8.7i0.1 3.4i0.2 0.13i.02 5/21 415 29.4i1.8 11.4i0.4 7.1iO.4 0.24i.02 5/25 500 28.5:2.1 16.6:0.4 13.111.1 0.46:.01 5/29 569 31.7i2.2 20.5:0.6 24.5:1.7 0.77:.03 6/3 677 47.9:3.3* 26.7i0.8* 59.1i4.4 1.23i.06* 6/7 797 45.0:2.8 34.3i0.7 96.8i6.8* 2.08:.07* 6/12 923 34.1i1.9* 42.7i1.0* 111.6i7.1* 3.27:,10* 6/17 1023 28.4:1.7 41.1i1.6 124.0i8.6* 4.36:.15* 6/21 1137 33.5:1.8* 47.2i2.0* 172.8i12.8* 5.15:.16* 6/25 1220 28.5i1.8 55.0i1.4 173.3:12.4 6.07:.28 6/28 1296 23.911.9 59.7i1.6 150.3i10.4 6.29:.33 7/2 1433 23.9i1.3 75.3il.8 177.6:13.2 7.43:.30 7/5 1543 24.7:1.5 76.5i2.5 165.4:12.1 6.69:.24 33 Table 6. ANOVA results for log 10 transformed oat stem density, plant wet weight and height, and per-stem weight for 13 sampling days in l974. Source of df Stems/ft2 Height Height/it2 Weight/stem VariatTon (N0.) (cm) (grams) (grams + 1) Crop (C) 2 .645**a .377** 1.61** .171** Date (0) 12 .596** 7.790** 33.837** 8.756** CO 24 .O69** .009* .073** .021 errorb .020(1129) .005(1129) .O31(1127) .OO8(1127) Total 1169 * .01 < P < .05 ** P < .01 Entries are mean squares and significance level b Error mean squares are followed by associated degrees of freedom. 34 lowest (Fig. 6). The source of the crop by date interaction is Clearly seen to be the response of pure oats, which did not display as rapid a decline in density after the peak as did the mixed stands. This may be due to the application of 2,4-D to the pure oats at 680 OD (arrow). The resulting reduction in the weed population may have allowed a higher oat density to persist. The discrepancy between stem density in OW and the other fields early in the season is due to a higher initial seed density. Oats were seeded at a rate of 2 bu/a, which gives 19.1 seeds/ftz. Wheat seeds are smaller, and 2 bu/a of a 50:50 mix of wheat and oats gives 26.1 seeds/ft2 (based on data from Hildebrand and Copeland 1975). These seed- ing rates are located along the vertical axis in Fig. 6 with the symbols 0 and ON. These seeding rates would contribute to the observed differ- ence in initial stem densities since for the OW treatment the oat and wheat stems were not separated. Plants were consistently taller in the OH field (Fig. 7) and were usually shorter in the pure oats. The growth appeared to take place in two phases: an initial growth of a steadily declining rate to a plateau around 1000 00, followed by a second increase in height, probably due to heading. The slight crop by date interaction (Table 6) is probably due to the increased difference between the treatments in this second phase of growth. Average wet weights (log10 transformation) are plotted in Fig. 8. No curves were drawn through the points because the treatment differences are small on the logarithmic scale. In general, OW produced the largest plants and pure oats the smallest, but on certain days these relationships 35 .Aw_mum mop mwxm PmngLm>v .mcowum>gmmno om mo :mwcms 0;» mm ucwoa comm .enmp cw muopa poucmswgmaxm 05p cw mumo mo xuwmcmu 50pm .o mcamwu moN¢ A 06 000. 00m . 000 00¢ F P P L « 3.22% d {LEN a r o Id I On mhzwc‘hdwmh ald/SWBLS 36 .Ampmom mo_ mwxa Fmowugm>v .mcowum>gmmno om 4o :mwuos asp we pcwoa gum“ .emmp cw myopa Pmucwewgqum ms» cw mama mo pgmwm; asap; .5 mgamwu to”? A Do Aunxw_ Aunmw_ nxumw nunxw Aunxm F D b 26:18--.. . 4 .1 I- a -111 s o a \ . mEmEfiE .om (W0) .LH913H .LNV'ld 37 .Ampmom mop mwxe Fmowusm>v .mcowgm>gmmao om Co :m_ums mzp m? pcwoa comm .enmp cw myopa Foucwewgmaxm 0;“ cm mpmo Co unmwwz pm: .m wszmwu mom¢ A Go 8.9 com. 2.5 0.8 on» _ m hm O 3010 4 <01". 6 m 014 a o 0. 352.2%» M 4 m. T w Mom m H a .mw o "nxu_ m a a a u m w a wow .08 (aw/6) .LHOIBM 1.3M 38 did not hold. These discrepancies are the source of the crop by date interaction (Table 6). Wet weight per stem was very similar for DA and OW (Fig. 9), but increased at a significantly lower rate for pure oats as the season progressed. Foliage samples in 1975 were only collected at the time of peak CLB larval densities and at the end of the season. Combining those samples collected during similar periods in 1974 with the 1975 data provides interesting information about the repeatability of the growth responses in different years. Table 7 tabulates the average responses for the two years. ANOVA's were calculated for each variable (trans- formed to logarithms) (Table 8). Averages were substituted for single missing values, as before, and the error degrees of freedom were corres~ pondingly adjusted. It is clear that second and third order interactions between the factors occurred for some of the growth parameters. To examine the nature of these interactions and to compare specific crOp effects, the average transformed values for each crop are given in Table 9 separated by year and sampling period. Means are compared using Duncan's new multiple range (NMR) test (Steel and Torrie 1960), .05 level. It can be seen (Table 9) that OW had the highest stem density in 1974 and the lowest in 1975. The overall result appeared in the ANOVA (Table 8) as a lack of significant effects due to crop. By examining the crop x year interaction it was thus revealed that planting wheat with the oats djg_affect the oat stem density, but this effect varied from year to year and was thus masked in the overall average. 39 .Ampmom mop mwxm F60mugm>v .mcowum>smmno om 4o AP mapav :mmums mg» mm pcwoa zomu .enmp cm mgopq Fmpcwswgmaxm wcu cw mpmo mo swam emu pgmwmz um: .m weaned n-omw‘v An Aun. com. com. com com 00» F. . . , . I _ 36551 J .n 5 .m .m 30\<0 m...-:a\\\\u. 40 110...... m o nu.111.a.1111 . 6. mhzwihdmmh (1 + swam) WELSI'LM 1.3M 4O Rpm.wn¢.m km.F~Ho.moP R~.Nwm.mm R¢.Fwn.mp zo PmmF vcm om.wmw.m P.¢Pwm.mop m.Pw¢.mm m.Fwo.¢N <0 Pomp new om.wmo.m m.prm.mmp m.Pw¢.nm o.FwN.oN o Pom. new mp.me.P N.¢wm.mm w.ch.FN «.mwm.mm 3o m¢o_ xmma No.0Nm._ w.NwN.N¢ N.OHN.NN F.Nwm.Fm <0 mqop xmma mm.w¢m._ o.¢wN.P¢ m.owo.mm N.Nwo._m o mqu xmma a dN.owmm.o F.NFH¢.mmF m.mwm.mu m.Fw~.¢N 30 mqmfi ecu mm.oflm_.~ w.crwm.n¢~ N.NwN.FN N.Pflm.om cmmao 40 once chww vcm we?“ Pm>smp xmma um macmspmmsv ppm toe Aom n 2 .mm w.mv mgwngmLma zpzogm ace .5 mFQmH 41 Table 8. ANOVA results for oat growth parameters (log10 transformed) at the time of peak larval densities and at the end of the season in 1974 and 1975. Source of Variation df Stems/ft2 Height Weight/ft2 Weight/stem (N0.) (cm) (grams) (grams + 1) Crop (C) 2 .0166a .0076 .1388* .0504** Sampling Period (P) 1 1.8674** 13.915** 15.724** 15.767** Year (Y) 1 .7970** 1.2703** 4.6649** .6821** CP 2 .0618 .0161* .0125 .0773** CY 2 .2097** .0794** .6533** .0626** PY 1 .1325* .4184** l.9817** .3826** CPY 2 .0036 .O367** .1412* .O720** Errorb ‘ .0234(347) .0052(347) .O368(346) .OO67(346) * .01 < P < .05 ** P < .01 a . . . . Entr1es are mean squares and s1gn1f1cance level. b Error mean squares are followed by associated degrees of freedom. 42 Table 9. Comparison of average growth responses for the three crops at the time of peak CLB larval densities and at the end of the season in 1974 and 1975. Crop* Parameter Year Period S2 n 0 0A 0W Stem/ft2 l974 Peak .0234 30 1.5431a 1.5724ab 1.6278b log1o(n) 1974 End 1.4070a 1.3556a 1.4335a 1975 Peak 1.4615a 1.4726a 1.4119a 1975 End 1.4070b 1.3556b 1.2692a Height 1974 Peak .0052 30 1.5557b 1.4847a 1.5320b log]0(cm) 1974 End 1.8176a 1.8439a 1.8860b l975 Peak 1.3470a 1.3496a 1.3148a 1975 End 1.8240b 1.8219b 1.7497a Weight/ft2 1974 Peak .0368 30 1.9577a 1.8674a 1.9577a log]o(g) 1974 End 2.1277a 2.1988ab 2.2598b 1975 Peak 1.5681b 1.6528b 1.4295a 1975 End 2.1880b 2.1702b 1.9912a Weight/stem 1974 Peak .0067 30 .5582b .4743a .4974a log]O(g+l) 1974 End .7986a .9034b .8879b 1975 Peak .3585b .3977b .3169a 1975 End .8858c .7975a .8410b * Means on the same line followed by the same letter are not significantly different at the .05 level (Duncan's new multiple range test). 43 Similarly, the effect of the companion crops on oat height varies both from year to year and between sampling periods within a year (crop x year and crop x sampling period, interactions, Table 8), even though crop type itself was not a significant factor. Furthermore, there was a 3-way interaction of crop x year x period. In Table 9, the average oat heights are separated by sampling periods and years to examine these effects. It is shown there that OW produced the tallest plants at the end of the season in 1974 and the Shortest in 1975. 0A produced the shortest plants at the time of peak larval densities in 1974, but all treatments were similar at the same period in 1975. The companion planting treatments were found to significantly influence oat wet weight (Table 8). But this effect also varied in the two years Crop x year interaction significant), and the interaction itself varied between the sampling periods within a year(crop x year x period). In Table 9 it can be seen that all treatments had similar weights at peak larval time in 1974, but in l975 OW had lower weights for the same period. At the end of the season, OW had produced heavier plants then 0 in 1974, but lighter plants in l975. Oat weight per stem was strongly affected by the planting treatment (Table 8). All interactions were also highly significant. In Table 9 the average per-stem weights are separated by sampling period and year. In 1974 pure oats had the highest weight per stem at peak larval time, but the lowest at the end of the season. The mixed crops were similar to one another throughout the year. In 1975 OW produced the lowest weight per stem for both sampling periods. OA had the highest weight per stem by the end of the season but was not different from pure oats earlier. 44 To summarize these effects of companion planting on oat growth, it can be said that at season's end in 1974, OW had a higher stem density, greater height, greater wet weight, and higher weight per stem than did pure oats. In 1975 these results were all exactly reversed (fewer and shorter stems, and lower weight per ft2 and per stem). For the end of the season in both years, the only effect of 0A was to produce higher weights per stem than did pure oats. The meaning of the interactions of the companion crop effects with the sampling date and year is not entirely clear. It is evident from Table 9 that the crop x year interaction is largely due to the ON com- bination. It is doubtful if crop plants (especially closely related species planted together) do not compete at least somewhat, and the relative competitive advantages of two species can be influenced by environmental conditions. Some of the factors that might determine the outcome of an interplanting are average, maximum, and minimum daily temperatures, soil water stores (influenced by precipitation and early spring snow cover) and hours of bright sunshine. There were no outstand- ing differences in the temperature patterns during the growing seasons of 1974 and 1975. However, 1974 received a more constant moisture supply (Fig. 10). In 1974 there was no snow accumulation during April at Gull Lake, and the total precipitation was 8.4 in. during April-May and 5.0 in. during June-July. In 1975 it was wetter in late spring (snow 4 in. deep in April and 12.4 in. of precipitation during April-May) and drier in early summer (4.4 in. of rain during June-July). It is impossible to conclude from just two "observations" (the final crop results and seasonal environmental states for the two years) just what environmental factors might be responsible for the observed interactions, but moisture is a 45 \é.‘ .\\\\;i :3 EE. 5: >5; \2 :5 F- g .1" 6'01;- (SBHDNI) NOIIVLIdIOBEM Figure 10. Total precipitation for April-July and maximum snow depth during March and April 1974 at Gull Lake for the two study years. Climatological data from U.S. Dept. Commerce, National Oceanic and Atmospheric Administration. 46 likely candidate. Under the 1975 conditions non-vernalized wheat may have been a more successful competitor with oats than it was in 1974. This does not explain, however, why the wheat would act as a competitor, reducing oat growth, in 1975, but would serve to enhance oat growth in 1974. A complex process is most likely at work here, and several years of data might be needed to unravel the interactions. It should be remembered that different varieties of wheat were used in the two years (Genesee in 1974 and Yorkstar in 1975), and differences in the agronomic characteristics of these varieties might account for the interaction of crop effects and year. This possibility should be carefully evaluated in subsequent work. The interactions between crop effects and sampling period within a year are perhaps even more difficult to interpret. Variation in environmental factors during the season could be the reason, or the differential use of the herbicide in one crop but not in the others may have altered the conditions enough between sampling periods to produce the observed interaction (the herbicide was applied near the day of peak larval densities in both years). Surface area of oat leaves can be estimated from a regression equation developed by Gage (1972): SA(mm2) = 52.15 + 0.034 (length (mm) )2 r = .88 Average leaf length and average surface areas calculated from the squared lengths are given for the uppermost three leaves in each treat- ment in Table 10 for 1974 and Table 11 for 1975. An analysis of variance (Table 12), using the method of unweighted means (Neter and Waserman 1974) 47 Table 10. Length and calculated surface area on July 4, 1974 of the upper three leaves of oat plants in the three treatments. Leaf Crop Length (mm)a n Surface area (mmz) Upper 0 135.9t5.6 50 731.8 OA 155.3:5.1 47 913.1 OW 167.9i5.7 50 1064.4 Second 0 243.5:4.5 49 2101.7 0A 261.1:4.8 47 2405.0 OW 282.4:4.5 50 2797.6 Third O 247.4:4.1 49 2161.1 0A 261.814.4 47 2412.5 OW 279.014.O 49 2723.5 S.E. >4 1+ 49 Table 12. ANOVA results for squared leaf lengths (log10 transformation) of the uppermost three leaves in 1975 using the method of unweighted means. Source of Variation df MS F Crop (C) 2 .0126 12.6** Leaf (L) 2 .4470 447.0** Year (Y) 1 .0018 1.8 C x L 4 .0010 1.0 C x Y 2 .0076 7.6** L x Y 2 .0021 2.1 C x L x Y 4 .0004 0.4 Error 687 .0010 ** p < .01 50 on the squared lengths (logarithmic transformation) showed significant treatment effects. Since leaf surface area is a linear function of the squared leaf length, differences in squared length imply differences in surface area. Multiple range comparisons of the means (Table 13) show that for the upper leaf, 0A had the largest surface area, and oats the smallest. For the second leaf, OA and OW had similar areas which were both larger than that of pure oats. For the third leaf, OW had the largest surface area, and pure oats again had the smallest. For the total of the three leaves, it was found that 0A and OW had similar total surface areas, and pure oats had a smaller total, at the .01 level of significance. Oat grain yield was calculated in 1974 from the head density and air-dry head weight (Table 14), using the following equation: yield = (heads/ftz) x (g/head) x (43,560 ftZ/a) x (.0022 lb/g) x (.0313 bu/lb) = bu/a where the head weight is first converted to thrashed grain weight by the relation developed for 'Garry' oats by Jackman (l976): Thrashed weight (g) .066 + .755 (Head weight (g) ) r = .955 The head weights for each field were estimated from a single weighing of 100 heads. An estimate of the variance for weight/head (S2 = .2 g/ head) was taken from Jackman (1976). The differences between the yield components in the mixed stands and the pure oats were found to be signi- ficant. OW yielded 52% more than oats alone, and 0A give a 41% higher yield than pure oats in 1974. These are astonishing differences. A 51 Table 13. Mean squared leaf lengths (log10 transformation) of the uppermost three leaves in the three treatments in 1975. Estimated variance = .001 with 687 df. -——————2 Leaf Crop log10 X Upper 0 4.2805 a 0A 4.4025 b OW 4.3975 c Second 0 4.7990 d 0A 4.8650 e OW 4.8635 e Third 0 4.7875 d 0A 4.8270 f OW 4.8535 e Means not followed by the same letter are significantly different at the .01 level (Duncan's new multiple range test). 52 Table 14. Oat head density (#/ft2), air-dry head weight (g) and estimated grain yeild (bu/a) for the three treatments in 1974 (i’: S.E.). Crop Head density Head weight Yield n = 30 n = 100 0 20.60:l.1l a 1.19:.045 a 61.1 0A 25.69:l.28 b 1.362.045 b 86.3 OW 24.94:l.30 b 1.52:.045 c 93.1 Means in the same column and not followed by the same letter are significantly different at the .05 level (Duncan's new multiple range test). 53 yield of more than 80 bu/a is very high.7 Since these values are cal- culated rather than actually measured as bushels of grain, they may be in error somewhat, but the differences between treatments are real. Since OA did not differ from pure oats in very many of the growth para- meters (Table 9), the yield difference is probably due to the greater leaf surface area of the oats grown with companion crOps. In order to reduce the error in estimating yield, in 1975 the head weight/ft2 was measured directly for ten ft2 per plot. This was done in all three fields at Gull Lake and also at Galien and Collins Rd. Yield is thus calculated more directly as: Yield (bu/a) = (g/ftz) x (43560 ftz/a) x (.0022 lb/g) x (.0313 bu/lb) where the head weight is first converted to grain weight by the regression presented above. An analysis of variance (logarithmic transformation) on head weight/ft2 showed that crop effects were highly significant (.01 level), but these effects interacted with field differences. Yield, in g/ft2 and bu/a, are given for each crop and field combination in Table 15. Also presented are the logarithmic means which were compared statis- tically. At Gull Lake there were no crop differences within fields, but there were field to field differences. In light of these data, the large 7The average yields for oats in 1972 for Kalamazoo Co. and for the entire state were 52.6 and 55.0 bu/a, respectively. (Michigan County Statistics, Field Crops, 1959-1972. Michigan Crop Report Service, Michigan Department of Agriculture 1974. 108 pp). 54 Table 15. Oat yield for the three treatments in several experi- mental fields in 1975. Yield . 2 2 F1e1d Crop g/ft 109(9/ft ) bu/a Gull Lake 9-10 0 21.2 1.302 c 49.5 OA 19.1 1.238 be 44.6 OW 23.1 1.316 c 53.9 8-10 0 13.5 1.078 b 31.6 OA 17.7 1.195 be 41.3 OH 13.3 1.064 b 31.1 5-55 0 18.9 1.228 bc 44.1 OA 20.7 1.252 bc 48.3 OW 20.6 1.228 bc 48.1 Galien O 16.8 1.208 be 39.3 OA 16.5 1.206 bc 38.6 OW 7.4 0.830 17.4 Collins Rd. 0 33.2 1.510 77.4 OA 34.0 1.504 79.2 OW 17.9 1.246 bc 41.8 Means of transformed data followed by the same letter are not significantly different at the .05 level (Duncan's NMR: n S 2 .038 with 135 df). 10, 55 differences observed in 1974 might be suspected of being due solely to the effects of the different fields the three treatments were planted in. However, several facts support the possibility that in 1974 the yields were truly greater as a result of the companion crops. First, OA and 0 produced greatly different yields but were planted in narrow, adjacent fields which had similar planting histories (Gage 1974: p. 29); thus the effect of field differences was minimized. Second, it has been shown that in 1975 the effect of the non-vernalized wheat was to reduce vegetative growth of the oats, and at both Galien and Collins Rd., the OW combination produced a significantly lower yield than the other plantings in the same fields. Yie1d should be well correlated with vegetative production, which in turn responds to growing conditions. Insufficient moisture during the critical period of heading (Fig. 10) could account for the lower yields observed in 1975, and with the very different rainfall pattern of 1974 oat growth and yield may have been enhanced by the companion crOps. In the final analysis, the conflicting results for the two years would prevent any definite conclusions about the value of the companion crops in increasing yields. Soil Moisture Soil water content, expressed as a percentage of oven-dry soil weight, is given in Table 16 for the three crops in three plots over a five-week period. These results are for the top one inch of soil, which is approximately the maximum depth for pupation of the CLB. An analysis of variance showed that the type of crop cover did not signifi- cantly affect the soil moisture in the top inch. Field to field 56 Table 16. Soil water content (% of dry soil weight) in the top inch of soil for the three crops over a five-week period in 1975. Field Crop Date 9-10 8-10 5-55 0 6/3 13.7 11.2 8.8 6/12 18.8 15.5 15.5 6/18 18.3 14.2 16.7 6/26‘ 9.1 8.2 8.1 7/1 5.6 2.3 3.5 7/9 1.5 0.4 0.6 OA 6/3 19.0 10.3 11.2 6/12 19.4 15.6 16.5 6/18 21.7 13.3 15.6 6/26 10.4 5.9 9.7 7/1 12.1 2.2 2.0 7/9 2.4 0.4 1.3 OW 6/3 14.6 10.4 15.5 6/12 18.7 17.0 15.7 6/18 26.8 13.9 16.5 6/26 10.4 7.2 9.6 7/1 8.9 1.1 2.4 7/9 3.4 0.4 0.8 57 differences and moisture changes with time were highly significant (p f .01). These results are plotted in Fig. 11. Soil Temperature In each treatment plot in each field, four thermocouples were buried in the soil: at two depths (.25 inch and one inch) and both within a grain row and between two rows. Temperatures were then recorded on three occasions (Table 17). An analysis of variance (Table 18) showed that all main effects (crops, depth, row placement, field and date) were highly significant. In addition, several two-way interactions were significant, including the interaction of crop type and field. The field effect is in part due to the influence of the time of day on soil temperatures, as all fields could not be measured simultaneously. Since an assumption of the ANOVA is additivity of variance components (Eisenhart 1947), when comparing crop effects it is possible to average observations over several other variables without altering the relationship between the means as long as these variables do not interact with the crop effects. For example, soil temperatures differed between depths, but this difference was the same in each crop. Thus the relationship between crops is not affected by averaging over depths. The advantage to this process is that the standard error, used for comparing means, is reduced by combining observations according to the equation: S.E. = VSz/n Means for the three crops are given in Table 19 separated by field, since the crop effects interacted with the field effects. It can be seen that 58 .mmaxu aogu mass» one gm>o mmmcm>m cm mw ucwoa comm .mnmp c? 0384 szu an mopmww moss» mo 50:? no» mg» cw mucosa weaumwoe pwom .FF mcaavm ”the ob m 22. pm mu ow m. 0.. A... 25.. B 6. .9 .o~ WWW ....I.....|...... .3 0.10 Ios 4111107.) .LNBLNOO HBLVM (mfiiam 1 59 .0000333 0>oam 30:3 0:0 .zaocmo 300:: 00033 :3 00030003 303300303500 330 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 333 3.33 3.33 3.33 3.33 3.33 _.3N 3.33 3.33 3.33 3 33. 3.33 3.34 3.33 3.33 3.33 3._3 3.33 3.33 3._N 3 33. 3.33 3.33 3.33 3.33 3.33 3.33 3.33 ~.3N 3.33 3 P 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3._~ 3 _ 33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 333 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3 33. 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3 33. 3.33 3.33 3.3m 3.33 3.33 3.33 3.43 3.33 3.33 3 3 3._3 o._3 3.33 3.33 3.33 3.33 3.NN 3.33 3.33 3 3 <3 3.33 3.33 3.33 3.33 3.33 3.33 3.~N 3._N 3.3m 4333 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.3m 3 3m. 3.33 3.33 3.33 4.33 3.33 3.33 3._N 3.33 3.33 3 3m. N.3N 3.33 3.33 3.33 _.33 3.33 3.33 4.33 3._N 333 3033303 3 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 3.33 333 333333 3 3 33-3 33-3 33-3 33-3 33-3 o_-3 33-3 33-3 33-3 330200333 383 330333 3830 3.3333 3 3333 N_ 0333 33303 3_033 333 0333 .mmmp .0303 33:0 00 0003000; Auov 30330030050» 3303-0>onm 0:0 Pwom .33 0—30» 60 Table 18. ANOVA results for soil temperature at Gull Lake in 1975. Source of variation df Mean Square F A Crop 2 31.51 ll.47** B Depth 1 207.50 63.48** C Row placement 1 31.26 9.56** 0 Field 2 516.35 157.95** E Date 2 1129.10 345.40** AB 2 3.24 0.99 AC 2 1.76 0.54 A0 4 65.51 20.04** AE 4 3.76 1.15 BC 1 .11 0.03 80 2 28.56 8.74** BE 2 19.15 5.86** CD 2 4.05 1.24 CE 2 3.65 1.12 DE 4 105.95 32.41** Error 74 3.27 Total 107 ** .01 "U IA 61 Table 19. Soil temperature (00) for the three treatments in each field at Gull Lake, averaged over depths, row placement and dates (n = 12, $2 = 3.27 with 74 df). Crop Field 0 0A 0W 9-10 25.76 a 24.68 a 24.79 a 8-10 33.43 b 31.48 a 34.01 b 5-55 24.89 a 31.24 b 30.40 b Crop means within a field followed by the same letter are not significantly different at the .05 level (Duncan's new multiple range test). 62 in one field the soil temperatures in all three crops were the same, in another field OA had lower temperatures, and in the third field pure oats had lower temperatures. Soil temperatures are influenced by a number of factors such as topography, soil type, soil moisture, air temperature, incident solar radiation, and air flow over the soil surface. The soil types in the experimental fields were all very similar, and the thermocouples were all placed in level areas. It has already been shown that the soil moisture was not different under the three crop types. The remaining factors are all influenced by the crop cover. The air temperature under the crop canopy, one inch above the soil surface, was measured each time the soil temperatures were recorded (Table 17). An ANOVA did not reveal any significant differences in air temperature among treatments. Several components of oat growth which contribute to the quantity of ground cover have already been reported for field 9-10 in 1975 (Tables 9 and 13). These results showed that compared to pure oats, OW had a lower stem density, shorter plants, lower wet weights per ft2 and per stem, and greater leaf surface area. The only differences produced by 0A were greater wet weight per stem and greater leaf surface area. In spite of these differences, the soil temperatures among the crops in field 9-10 were not found to be different. None of these crop growth components were measured in the other fields, where soil temperature differences were found. These sets of data thus add nothing to an understanding of the relationships between soil temperatures and the nature of the crop cover. 63 Another measure of crop cover which was examined for each field is above-ground dry weight of foliage. These data were gathered to determine what the total plant nitrogen content was for each treatment, and how this nitrogen was distributed among the various groups of plants (oats, alfalfa, broadleaf weeds, and grassy weeds) in each field. Above-ground dry matter in these components for each crop in each field on each of five dates is given in Tables 20-23. An ANOVA on total dry matter/2 ftz, transformed to logarithms (Table 24), showed signifi- cant differences due to dates and fields, but no effect of crop type. Similar results were obtained for the oat and grassy weed components of dry matter. For the broadleaf weed component, significant effects due to fields and a crop x date interaction were found. This last inter- action is understandable in light of the application of a herbicide to the pure oat stands on June 16, 1975. The dry weights for each component, averaged over fields and crops (except for broadleaf weeds, which are separated by treatments) are plotted in Fig. 12 to show the trends for biomass accumulations. Total dry matter increased rather smoothly throughout the period of measurement. This increase was predominately due to the oats, while grasses and broadleaf weeds increased only slightly. In the pure oat fields, the broadleaf weed dry matter actually declined, and an unexplainable reversal of relative dry matter accumu- lation occurred in the broadleaf weeds in the ow and 0A plots on July 9 (last date). Since there were no differences between the experimental treatments with regard to plant biomass in any field, once more we are left without an explanation for the differing soil temperatures observed. Any error DRY MATTER g/2ii‘z I001 80‘ 60 201 64 TREATMENTS : O T43—T (MAL TCTHAL. :33 ()Vv ,’/. O a 0 fl 0 'O o 1000 Figure 12. 1200 1400 1600 1800 °o > 42°F Dry matter accumulation on experimental plots at Gull Lake in 1975. Each point is an average over nine plots (except broadleaf weeds, where each point is an average over three fields). 65 Table 20. Accumulation of total above-ground dry matter (g/2 ftz) at Gull Lake plots, 1975 (includes oats, alfalfa, broadleaf weeds and grassy weeds). Date Field Crop 6/12 6/18 6/25 7/2 7/9 9-10 0 25.74 30.04 37.24 89.86 104.56 0A 31.70 56.34 59.83 89.54 105.88 OW 15.96 27.84 44.49 60.32 93.85 8-10 0 22.79 35.61 62.50 80.10 73.31 0A 27.13 55.94 80.96 118.82 89.77 CH 44.93 38.81 88.68 46.87 124.81 5-55 0 21.90 29.69 34.99 65.99 104.02 0A 22.44 29.09 53.41 51.01 61.71 OW 14.50 31.83 37.64 59.03 82.26 lat 11 F11 66 Table 21. Accumulation of above-ground oat dry matter (g/2 ftz) at Gull Lake, 1975. Date Field Crop 6/12 6/18 6/25 7/2 7/9 9-10 0 6 18 16 36 22.29 63 43 83 38 0A 8 50 29 29 25.20 50 09 81 38 OW 6 93 ll 19 21.13 33 O8 59 56 8-10 0 ll 20 21 75 48 50 60 71 59 50 0A 12.39 39.85 46.56 97.22 54.97 OW 37.05 31.69 57.96 40.38 93.11 5-55 0 13.15 25.11 24.11 53.09 98.24 0A 12.38 22.58 44.02 38.45 44.44 67 Table 22. Accumulation of above-ground grass dry matter (g/2 ft2) at Gull Lake, 1975. Date Field Crop 6/12 6/18 6/25 7/2 7/9 9-10 0 15.03 6.62 7.74 22.10 17.18 0A 14.17 16.00 21.50 24.11 18.46 ON 7.05 6.90 19.50 13.73 15.80 8-10 0 3.51 6.49 6.32 9.05 6.60 GA 6.91 6.94 16.99 10.13 21.91 0W 4.53 0.81 21.17 1.59 21.52 5-55 0 3.39 2.77 9.52 9.20 1.23 GA 6.52 3.02 5.62 8.14 15.70 0W 6.25 9.32 7.43 8.12 11.72 68 Table 23. Accumulation of above-ground broadleaf weed dry matter (g/2 ft2) at Gull Lake, 1975. Date Field Crop 6/12 6/18 6/25 7/2 7/9 9-10 0 4.53 7.06 7.21 4.33 4.00 GA 8.72 9.99 10.21 15.34 5.52 ON 1.98 9.76 3.86 13.51 18.49 8-10 0 8.08 7.37 7.78 10.34 7.21 0A 3.41 7.29 8.56 8.53 11.89 CH 3.35 6.31 9.55 4.90 10.18 5-55 0 5.36 1.81 1.36 3.70 4.55 CA 2.42 3.19 2.88 4.15 1.46 0W 1.33 2.42 3.92 3.76 15.55 69 Table 24. ANOVA results (log10 transformation) for dry matter compo- nents in the Gull Lake plots, 1975. Entries are mean squares and significance level. Source of Needs Variation df oats broad leaf grassy Total Crop 2 .0144 .0133 .1972 .0298 Field 2 .1728** .6648** .5502** .0955** Date 4 .8725** .1053 .2226* .4822** Crop x date 8 .0177 .1067* .0855 .0124 Error 28 .0231 .0395 .0729 .0115 Total 44 * .01 < p < .05 ** p < .01 70 in placing true thermocouples at the proper depths would result in add- itional variation in the recorded soil temperatures since there is a strong effect of depth on this variable. However, there is no evidence that this occurred, and the problem remains unresolved. Plant Nitrogen Nitrogen content of above-ground dry matter was measured (Table 25) to determine if the total nitrogen per unit ground area, the portion of total nitrogen in each plant component, or the percent nitrogen in oats differed among the treatments. Total plant-contained nitrogen per 2 ft2 is given in Table 26. An ANOVA showed significant effects due to crop treatment, field, date, and an interaction between crop and field (Table 27). To compare specific crop effects, the crop means must be separated by field, since crop and field interact. An average over the five dates is given in the last column of Table 26. In fie1d 9-10, 0A had a greater total nitrogen level than ow, averaged over the five dates. Pure oats was not different from either mixed crop. In fie1d 8-10, 0A had a greater nitrogen level than pure oats, but ow was not different from either of these. In fie1d 5-55 all crops had similar average nitrogen levels. The data are plotted in Fig. 13. The effect of the alfalfa was not as great as might be expected from the many reports on the beneficial effects of leguminous crops (see p. 5). Since the fields were fertilized at planting, there may have been an excess of nitrogen in all treatments which the plants could not use because they were limited by some other factor. It is clear from 71 53'") ti-K) I20 5955 ' eoo m ‘7 l oooo moo 13001000 I400 1130010001400 1900 °D > 42°F Figure 13. Total plant-contained nitrogen (mg/2 ft2) in each plot over oD>42. Each point is a single sample; lines were fitted by least-squares. 72 Table 25. Nitrogen content by date (g/2 ft2) for all plant components (0 = oats, G = grasses, BLN = broadleaf weeds, A = alfalfa) in the Gull Lake plots, l975. Sampling_Date Field Crop Component 6/12 6/18 5/25 7/2 7/9 9-10 0 O .211 .359 .447 .872 1.212 G .474 .151 .151 .392 .286 BLN .123 .120 .144 .109 .076 0A 0 .248 .736 .447 .788 1.419 G .427 .393 .435 .455 .279 BLW .191 .183 .145 .301 .068 A .007 .020 .046 .0 .009 ON 0* .124 .278 .371 .488 .723 G .155 .153 .393 .263 .227 BLN .049 .017 .060 .267 .271 8-10 0 0 .432 .464 .873 .810 .746 G .132 .158 .137 .166 .108 BLN .219 .128 .136 .179 .141 0A 0 .391 .962 .854 .190 .717 G .226 .190 .346 .181 .321 BLW .117 .181 .191 .203 .180 A .139 .056 .244 .081 .024 ON 0* .976 .641 .854 .527 1.168 G .129 .018 .401 .021 .304 BLW .072 .099 .139 .092 .147 73 Table 25 (cont'd.). Sampling Date Field Crop Component 6/12 6/18 6/25 7/2 7/9 5-55 0 0 .388 .601 .418 .825 1.075 G .104 .069 .194 .194 .016 BLW .118 .045 .026 .082 .086 0A 0 .433 .464 .658 .486 .486 G .214 .077 .118 .150 .244 BLN .059 .071 .053 .077 .029 A .024 .005 .014 .005 .002 ON 0* .178 .372 .435 .572 .591 G .164 .194 .163 .142 .213 BLW .030 .040 .059 .050 .229 Includes the non-vernalized wheat, which was not separated from the oats. 74 Table 26. Total above-ground plant nitrogen (g/2 ftz) at Gull Lake in 1975. Date Field Crop 6/12 6/18 6/25 7/2 7/9 All dates* 9-10 0 .808 .624 .742 1.372 1.574 1.024 bcd 0A .873 1.332 1.074 1.534 1.776 1.318 de ow .328 .448 .823 1.017 1.220 .767 ab 8-10 0 .783 .750 1.146 1.155 .995 .966 abc 0A .873 1.389 1.634 1.654 1.242 1.358 e 0w 1.177 .758 1.394 .639 1.619 1.117 cde 5-55 0 .610 .715 .638 1.101 1.176 .848 abc 0A .730 .617 .842 .718 .761 .734 ab ow .372 .607 .657 .765 1.032 .687 a Means in this column followed by the same letter are not significantly different at the 5% level (Duncan's new multiple range test. s2 = .0454 with 16 df, n = 5). 75 Po. v a .3. mo. v a v Po. Am_v Amvm_o. -- -- Aevemo. -- .Amvmo_. .Amvmoo. cometpwe Hematma e .m=< .mumm; “no Aeev Ammvemo. -- .Amvkm_. -- .lfievwwe.e Amvmmo. «.ANVNee. camocpwc N “no Ae_v AmVF_ -- -- -- Aevmm .ANVFm -- mcch_< heev Ammvmm -- Amvmo -- A¢V_m .Amvm_F Amvm memo: campumOLm Aeev Awmvkm -- Amvmm -- vfievmmm .kfimvamop A~V_NP mammmtw Aeev Amwvom_ -- Amvmm -- vsfievfimo ..Amvom__ Amvkmp mpmo ”cw quou mo pcmogma Acev Am_vm¢o. Amvpmo. Amvpoo. *54me_. ..Aevmme. *IANVoFo. vsamvoom. Ampcm\mv _mpop Payee coca” axe oxu axu mung upm_a aocu mpnmwtm> cowumwcm> mo muczom .pm>o_ 60:80wmwcmwm new Aeocmmcm 4o mwmcmmuv .mmcmzcm 2862 age mmwcpcm .mucwcoasou cmmogu_: ucwpa ccaogmum>onm Lo; mppammg <>oz< .NN mpnmh 76 the ANOVA results that under the experimental conditions, field differ- ences were greater than the effect of the crop type on the total plant nitrogen. It is likely that differences between sites within a field were also very great. Local differences in nitrogen level are probably influenced by slope, depressions or crests in the topography, and soil moisture. Because of large variations, a single sample from each plot per day may have been inadequate to reveal the overall crop effects. Also, it might have been incorrect to assume that the total above-ground plant nitrogen reflects the level of available soil nitrogen. A portion of the nitrogen contributed by the alfalfa might have been unavailable to other plants, or if absorbed, might not have been translocated to the above-ground plant parts, or, as mentioned above, simply might have been in excess of what the plants could use. Milthorpe and Moorby (1974) state that the great season to season variation in yield response of a crop at a particuliar site to levels of fertilization suggests that a number of intricately interrelated components ("such as initial concen- trations and rates of release and immobilization in the soil, rates of root expansion and of photosynthesis in the various leaf layers, the amount of transport from older leaves to the current growing organs, and so on") influence the end result. This complex of interactions may further explain the different responses of oats to companion planting in the two study years. Of the total plant nitrogen, the proportion which was contained in each plant type (oats, grasses and broadleaf weeds) did not vary among the crop treatments (Table 27). These proportions did vary significantly 77 from field to field, and the proportions found in the oats and grass components changed with time (increase for oats and decrease for grasses) (Fig. 14). The actual nitrogen content of the oat plant material (percentage of dry weight) had significant variance components associated with treatments, dates, and the interaction of these two factors (Table 27). A plot of the crop averages (Fig. 15) shows that most of the crop effect and the crop x date interaction is due to the low value for ON on the first sampling date. This may be explained in terms of the different nitrogen content of wheat and oat leaves. Jackman (1976, Fig. 7) shows that wheat leaves at 1000 0D > 42 have a lower percentage of nitrogen than do oat leaves. Early in the season, while wheat plants (which were not separated from oat plants) constituted a sizable proportion of the plant biomass, the overall nitrogen content in the ON samples would thus be lower than if samples consisted solely of oats. Other than this one anomaly, the percent nitrogen in the oat foliage decreased with time and did not differ between treatments. Since the expected effect of the alfalfa was the addition of nitrogen to the soil, these results are consistent with Jackman's (1976) findings that fertilizer levels do not affect the nitrogen content of oat leaves. This is further corroborated by Milthorpe and Moorby (1974, p. 162). The percentage of nitrogen in oat heads on August 4, 1975 differed among treatments and fields (Table 28). The mean nitrogen content of oat heads in the pure stands (1.87%) was greater than the means for GA (1.70%) and ON (1.68%) at the 5% level of significance (Duncan's new multiple range test). 7. OF TOTAL 80' 60‘ 40« 20‘ 1000 78 OATS GRASSES A fi fi v A fi 1200 1400 1600 1300 °o > 42°F Figure 14. Percentages of total above-ground plant nitrogen per 2 ft2 in each plant component, by date. Each point is a mean of nine observations (except for alfalfa, where each point is a mean of three observations). 79 .cowgmw>mu ngmucmpm F n m? can _muPpr> .mcowum> 1meno macs» mo come a m? ucwoa gumm .mxmu Pmcm>mm :o pcmsummcu comm com mmmsown umo ucaocm1m>onm Page“ yo Apgmwmz xgu may cmmoguw: gcwugma .mp mcsmvu h. ON¢ A 00 81». one. 8.... 08. 08. RM. gu.~ (111011111 Mp M1) 111390111110 % 80 Table 28. Percent nitrogen (by weight) of oat heads on August 4, 1975 at Gull Lake. Samples were two linear feet of grain. Field Crop Subsample % nitrogen 9-10 0 1 1.654 2 2.024 0A 1 1.944 2 1.804 ow 1 1.794 2 1.614 8-10 0 1 2.134 2 2.024 0A 1 1.654 2 1.714 ow 1 1.814 2 1.744 5-55 0 1 1.634 2 1.764 0A 1 1.614 2 1.484 0w 1 1.614 2 1.474 81 Insect Response to Companion Cropping CLB Density. CLB densities estimated from the foliage samples collected on 13 dates in 1974 are given in Tables 29-31. As with the measures of plant growth, the variances of these density estimates are positively correlated with their respective means (Fig. 16). At low densities, the distribution of insects approaches a random, or Poisson distribution, with the variance very near the mean. As the density increases, the variance soon becomes much greater than the mean, in- dicating an increasing aggregation of the insects. The relationship between the variance and the mean appears to be the same for eggs and larvae, and can be expressed by the regression: log 52 = .245 + 1.247 log i r2 = .97 n = 65 standard error = 0.154 The development of the insect populations in each of the three fields appeared to proceed at approximately the same rate. That is, events such as first and last observed larva, and peak egg and larval densities occurred on nearly the same dates in each field. To compare populations in the fields, the peak egg and larval densities [10910 (x + 1)], were compared using Duncan's new multiple range test (Table 32). The variance estimates were pooled from the within-treatment variances of the transformed data. All treatment means are different from one another at the .01 level of significance. Thus, 0A had the highest and ow had the lowest peak egg and larval densities. Gage (1974) has shown that peak egg and larval densities are highly correlated with the total seasonal population of these stages. VARIANCE 82 ICX31 0 Eggs :1 Larvae ' fi 0 «no , X. 0 O 10‘ ’11 ' X C x . O “'1 d- '* ' s x. ' I: .X. x O .i O O O O .0 X. 0 O X X .14 ’ 9: fl 0 IN j, . . .4 . . IN .I 1 1C) 2C) MEAN DENSITY Figure 16. The relationship between the variance and mean for CLB egg and larval densities per ft2 in oats at Gull Lake in 1974 (both axis log scale). Each point is a mean of 30 observations. 83 mp1m upmvw wo qo.wmo. mo.wmo. o mo.wmo. o eo.wmo. memp m\N mo.nPP. mo.wmo. «0.450. c o ep.nom. Amp_ N\N mm.nmo._ 8F.wmm. OF.HoN. eo.Hmo. mo.nmo. mN.HNm.F meoP mm\o mm.nmo.P mo.umo. No.np_. mp.wmm. o_.nmm. mm.nmm. mam mN\m NN.4_©.P o_.wfim. mo.flmm. e_.nmm. mo.wo~. em.nmm._ mum _N\m Fm.HN¢.m mm.n~m. 5F.H~N. 5F.4Nm. 5F.nom. mn.wee.m mmw nP\m mm.fiom.e N_.HNF. om.nmm. me.ne~.~ mp.nom.F em.nmw.m owe N_\m mm.wmm.m o no.we_. _~.noo._ Ne.nmp.m mm.n~o.o mom ~\o om.noe.P o o mF.HmN. wN.we_.P om.no_.m Rom m\m moo.noo. o o o moo.woo. mm.nm_.u mme mm\m mo.nmo. o o o mo.nmo. mm.umm.~ Dee m~\m o o o o c N_.Hem. mum F~\m o o o o o mo.nm~. m_m NF\m 55>254 _5304 >H HHH HH H mmmm mexoo mama mm<4 “om u c .mm u my .eNmP cw po_a muwo mesa ecu cw mmmwpm wmwp m4u mo Ampm\ocv zpwmcmo .mN mpnoh 0 0 0 0 0 80.800. 000_ 8\0 80.880. 00.800. 00.800. 0 0 N_.800. 08p. «\0 8P.8_0. 0F.888. 00.800. 80.800. 0 0_.808. 0008 8~\8 00.800._ 00.800. 80.800. 08.808. 00.800. 00.800. 880 8~\8 N0.888.N _N.800. 0_.800. 0_.808. 8_.880. 8F.80N.F 8N0 _N\8 08.88P.0 08.800. 0_.880. 00.880. 00.800._ 80.800.0 800 NF\8 N0._800.8 80.88_._ 00.888.” 80.888.N 00.800.N .8.80_.8 080 NP\8 a. 00.888N.8 0 00.880. 80.800.~ 80.808.8 00.80N.PF 088 “\8 no 00.800.P 0 0 N_.880. 00.880.P 8N._80P.0_ N88 0\8 00.880. 0 0 80.880. 80.800. 00._800.0P 000 0N\8 0 0 0 0 0 80.HNN.8 088 8~\8 0 0 0 0 0 00.880._ 800 _N\8 0 0 0 0 0 00.808. 000 n_\8 88>888 F8800 >0 000 HH 0 8008 80x00 8880 m<0mzH 4<>m<4 A00 u c .mm H my .enmp cw mp1m 8_880 08 88_88P8 088; 8880880 8880 08 880888 8880 000 80 A008\000 0888080 .00 8_880 II C 0m .38 85 00 u c 8 0 0 0 0 0 0 00m_ 0\0 00.800. 00.800. 0 0 0 00.800. Ampfi «\0 0 0 0 0 0 00.800. 0000 mm\0 00.800. 0 0 00.800. 0 N_.800. 000 00\0 00.800. 00.800. 0 00.800. 0 00.800. 000 88Pm\0 00.800. 00.800. 00.800. 00.800. 00.800. 00.800. 000 np\0 0F.800. 00.800. 00.80P. 00.800. 00.800. 00.8F0.F 000 8NF\0 0F.8_0. 0 00.800. 0F.8P0. «P.800. 00.800._ 000 0\0 00.800. 0 0 00.800. 00.800. 00.800.N 000 «80\0 000.800. 0 0 0 000.800. 00.800.? 000 0N\0 0 0 0 0 0 00.80F.P 000 00\0 0 0 0 0 0 0_.800. 000 ~N\0 0 0 0 0 0 00.8PF. 0P0 0_\0 88>L80 F8800 >H HHH HH H 8000 00800 8880 mm<0 A00 u c .00 8 my .0000 :0 PF10 8F880 :8 88803 0883 8880800 8880 :8 880888 wwwp 000 we A080\ocv 0808080 .00 8F080 86 Table 32. Comparison of the 1974 peak CLB egg and larval densities [log.IO (no/ft2 + 1)] in the experimental fields. EGGS LARVAE Field Crop Y’ 52 1' 52 9-16 0 .890 .075 .674 .115 9-17 CA 1.198 .045 .946 .075 9-11 ON .522 .053 .258 .057 Pooled $2 .058 .082 Degrees of freedom 86 87 87 The density of pupal cells in the soil and of adult CLBs emerging from these cells in 1974 are given in Table 33. These data were obtain- ed from soil samples and emergence traps, respectively. It was assumed that these stages were also "contagiously" distributed, so the densities were transformed to common logs prior to comparison. The density of pupae was lower in ow than in pure oats and 0A, and the density of emerging adults was higher in OA than in the other two treatments. In 1975, square foot samples were collected at Gull Lake on only one sampling date, near the estimated time of peak larval densities (June 13). These densities were extremely low (Table 34); hence in making comparisons it was assumed that the data followed the Poisson distribution, and counts were transformed by y' = «y—i—T5'(Mendenhall 1968, p. 208) prior to analysis. On this date there was no difference in egg density among the plots (Table 35), but the larval density in OH was significantly lower than in the other two crops. The rank order of the densities was also the same in both years (0A highest, 0 intermediate, 0N lowest). As a further test of density differences, the weekly sweepnet catch, converted to larvae/ft2 by the method of Ruesink and Haynes (l973)*, was used to estimate the total larval population for the season. The density for each field was plotted over physiological time (00 > 48) and the area under this curve was calculated. This area, when divided by the * This method has inherent biases in that it disproportionately picks up large larvae and misses small larvae. See Fulton (1975) and Logan (1976) for further discussion of sweepnet sampling for the CLB. 88 A022 8.0880000 800888880088 80 _8>8F F0. 88 888080 :80» 888— 8.20 F8>8P 00. 88 888080 0808 8888880 .1 08 0P 0 000.8000. 00.0800.0 88000.80NP.0 80.0800.0 30 p010 8000.8000. F0.0800.0 000.8000.0 00.0800.0 <0 0P10 000.8_00. P0.N800.N 000.8000.0 00.0800.0 0 0_10 AF + xv 000 ~00\00 AP + xv 00p 080 »\00 0080 00880 800000 00000 .000_ :8 88_:08 885208 00808858 008 8FP88 P8000 000 80 A.0.8 8 x0 888888080 .00 8F08H 89 80.880. 0 0 80.880. 0 80.880. 30 .02 08-0 08.880. 80.888. 80.880. 0 80.880. 80.880. 00 .88 08-0 80.888. 80.880. 80.880. 80.880. 0 80.880. 0 .88: 08-0 888888 88888 >8 888 88 8 8008 0080 88888 mm<0 .A00 0 z ..0.m + 00 8808 .88 8080 00 8888 8880 88 880888 8888 880 80 88888000 8888080 .08 88888 90 Table 35. Comparison of CLB egg and total larval densities (/y + .5 transformation) in field 9-10 on June 13, 1975. EGGS LARVAE Crop 2' 32 i' 82 o .724 a .009 .776 b .032 0A .742 a .017 .817 b .068 ow .724 a .009 .630 a .141 Pooled $2 .012 (87) .080 (87) Means in the same column and followed by the same letter are not significantly different at the 5% level (Duncan's NMR). 91 larval developmental time, yields an estimate of the total production of median-aged larvae (Southwood 1966, p. 279). The areas under the density curves were approximated using the trapezoidal rule, and a larval developmental period of 240 0D (>480F) was assumed (Tummala et a1. 1975). The seasonal larval populations calculated in this manner (Table 36) were then analyzed by ANOVA. Since it would be difficult to guess how estimates of seasonal densities calculated by this complex procedure would be distributed with repeated sampling, the analysis was simply done on the raw estimates, without transformation. The mean seasonal density of larvae in the pure oat plots was found to be higher than the densities in OA and ow. This result therefore conflicts with the previous finding from foliage samples collected near the time of peak larval density that 0 and 0A were not different from each other. The density of pupal cells and emerging adult CLBs at Gull Lake in 1975 are given in Table 37. Analyses of variance (¢§—:_T5 transfor- mation) showed that no differences attributable to the companion crops were evident. At Galien on June 10, 1975, no significant differences were found between treatments with respect to egg and larval densities. However, on July 29, a lower density of pupal cells was found in the ON plot. (Table 38). This indicates a higher mortality of larvae in ON between June 10 and the time of pupation. The establishment of oats was not very good in any of these plots due to competition from weeds, especially in ON where the seeding rate of oats was half of that in the other two plots (50:50 with wheat). Recall that the yield of grain was also lower 92 Table 36. Larval densities (no/ftz) estimated from sweepnet catches and estimated seasonal population for each plot in 1975. See text for methods of calculation. LARVAE/FTZ Field Date 0 0A 0N 9-10 5/27 .0 .0 .0 6/9 .0357 .0306 .0255 6/16 .1020 .1020 .0867 6/23 .0 .0204 .0153 6/30 .0 .0 .0 Seasonal .089 .101 .082 8-10 5/27 .0 .0 .0 6/2 .1709 .1428 -- 6/9 .1530 .1785 .1428 6/16 .1479 .0714 .1785 6/23 .0765 .0357 .0255 6/30 .0 .0 .0051 Seasonal .284 .203 .223 5-55 5/27 .0 .0 .0 6/2 .0102 .0153 .0357 6/9 .0306 .0102 .0051 6/16 .0969 .0306 .0816 6/23 .0357 .0306 .0 6/30 .0 .0 .0 Seasonal .108 .053 .068 MEAN SEASONAL .160* .119 .124 Significantly higher than 0A and ON at .05 level (Duncan's NMR; 52 = .00029 with 8 df). 93 Table 37. CLB pupal cell and emerging adult densities at Gull Lake in 1975. (i i 3.0.) PUPAE ADULTS Field Crop no/% N2 no/yd2 n = 20 n = 10 9-10 0 .301.57 .20i.42 0A .35:.8l .50:.71 ow .20:.41 .10:.32 8-10 0 .10:.31 .30:.48 0A .15:.49 .10:.32 0w .25:.72 .40:.70 5-55 0 .350.67 .10:.32 0A .20:.41 .20:.63 ow .10:.45 .30:.48 All Fields 0 .25:.54 .20:.41 0A .23:.60 .27:.58 0W .180.54 .27i.52 94 Table 38. Mean egg and larval densities on June 10, 1975 and mean pupal densities on July 29 at Galien, and error mean squares from analyses of variance. Crop Eggs/ft2 Larvae/ft2 Pupae/I/zyd2 n = 20 n = 20 n = 8 0 18.28 6.19 8.25 0A 22.32 8.13 7.75 ow 18.45 7.05 2.25* MSE (df) 59.84 (57) 21.87 (57) 24.40 (21) Significantly different at .05 level (Duncan's NMR). 95 in ON at Galien (Table 15). The lower larval survival in this plot may be related to increased dessication in a sparse canopy, greater predation where there were relatively more weeds to harbor predators, or greater feeding competition where an equal number of larvae were distributed among fewer stems. Unfortunately the Galien plots were not studied in sufficient detail to evaluate these possible factors. At Collins Road on June 27, the mean numbers of larvae per 50 sweeps with a sweepnet were found to be significantly different in each plot. Oats had the highest number (14.5), then 0A (9.0), then ON (4.2) (ANOVA MSE = 5.306, F = 19.84 with 2 and 9 df). Fulton (1975) has shown that very slight differences in the synchrony of two populations can re- sult in greatly different estimates of total seasonal population based on a sample from a single date, since densities build up and decline very rapidly. The weighted-mean instar (WMI) (Fulton 1975) may serve as an indicator of the maturity of a population, if the sample is not taken too near the end of the season. The NMI for 0, 0A, and OH were 3.87, 3.30, and 3.59, respectively. While these differences may be large enough to represent populations of different maturity, the plot with the median maturity (0N) had the lowest larval density. Thus it is unlikely that the observed density differences can be attributed to this effect, but represent different seasonal production. Parasitism of CLB Larvae Several independent estimates of percent parasitism by Tetrastichus julis can be made from the data collected at Gull Lake. Table 39 gives dissection results for large larvae hand-picked in 96 Table 39. Percent parasitism by I, juli§_of large CLB larvae collected for dissection in 1974. Numbers in parentheses are the number of larvae dissected. CROP Date 0 0A ON 6/12 53.8 (26) 61.5 (26) 33.3 (27) 6/18 37.5 (24) 44.4 (27) 23.1 (26) 6/21 37.9 (29) 22.2 (27) 26.1 (23) 6/25 19.2 (26) 10.3 (29) 23.1 (26) 6/28 34.8 (23) 51.9 (27) 53.9 (26) 7/2 88.5 (26) 69.6 (23) 66.7 (3) Weighted i 38.6 (154) 39.5 (159) 30.8 (131) 97 three experimental fields during the period June 12 through July 2, 1974. Approximately equal numbers were dissected from each field on any given date. The weighted mean percent parasitism rates do not differ significantly from field to field despite the wide range of larval densities. This indicates that I, 10115 females are probably limited by some factor such as searching capacity and were not satiated with hosts. As the density of hosts increases, hosts are easier to locate and more are parasitized. Another measure of larval parasitism is the proportion of CLB pupal cells in the soil which contain 1, juljsf. These data for 1974 are given in Table 40. These results also show no significant differences between treatments. The parasitism rates based on pupal cells are not comparable with those from hand-picked larvae because the latter does not take into consideration the interaction of CLB density with daily parasitism rates (every date was given approximately equal weight by the method of collect- ing about 25 larvae regardless of density). The parasitism estimates from soil samples more correctly reflect the seasonal impact of I, 10115 since the empty cells present at the end of the season represent an integration of daily dynamics. In 1975 soil sample estimates of seasonal larval parasitism were again made. Another estimate is also available from sweepnet collections of larvae. Table 41 summarizes these estimates. On each day of sweepnet *Empty cells can be classified by the size of the hole in the cell. CLB's leave large holes while I, julis leaves a small round hole and several exuviae within the cell. 98 Table 40. Percent parasitism of CLB larvae by I;_julis in 1974 as measured by examining pupal cells collected in soil samples. (15 15yd2 samples per field). Crop Total Cells Parasitized Cells % Parasitism O 51 28 54.9 0A 76 38 50.0 ON 10 6 60.0 99 Table 41. Estimates of larval parasitism at Gull Lake in 1975 based on sweepnet collections and soil samples. Numbers in parentheses are total larvae dissected to arrive at each estimate. % PARASITISM Crop Sweepnet Soil Samples 0 2.70(135) 16.7(12) 0A 5.95(84) 25.0(12) 0N 6.09(115) 37.5(16) 100 sampling, an equal sampling effort (200 sweeps) was made in each field. Higher densities, therefore, contributed more information to the estimate of parasitism. The number of parasitized larvae caught during the season divided by the total number of larvae caught thus gives a percent para- sitism which takes into account (integrates) the relative changes in larval density and parasitism rates. This estimate should thus be comparable to the estimate based on soil samples, which, as explained above, is also an integration of seasonal dynamics. It is obvious, however, that the two sampling methods produce very different estimates of seasonal larval parasitism. Several possibilities exist: (1) There is a differential mortality of parasitized and non-parasitized larvae between the time of exposure to sweepnet capture and the time of pupal cell formation. This would have to favor the survival of parasitized larvae in order to produce the observed discrepancy, and this is unlikely. (2) Unparasitized larvae are more subject to capture by the sweepnet than are parasitized larvae. This could be due to a different vertical dis- tribution on the plants, or to a different behavioral response to the approaching sweepnet, such as becoming dislodged more easily. (3) Since the sweepnet samples include some small larvae which have perhaps not yet been parasitized, this estimate would be lower than the soil sample estimate which represents the final outcome of larval parasitism. While this is probably of some importance, the known bias of sweepnet sampling for large larvae (Fulton 1975) would compensate somewhat for this effect. Evidence from samples taken on June 24, 1975 from another oat field (5-3) at Gull Lake also indicates that the sampling bias of the sweepnet for older larvae is insufficient to explain the discrepancy. In this example 101 three out of 38 fourth instar larvae (7.9%) collected by the sweepnet were parasitized while 19 out of 21 fourth instars (90.4%) hand-picked from foliage were parasitized. (4) The soil samples may differentially recover parasitized and unparasitized pupal cells. This possibility was tested and found to be not true (Appendix B). Thus possibility (2), a differential probability of capture for parasitized and unparasitized larvae, appears to be the most likely source of the observed discrepancy. This problem certainly deserves further investigation. In any case, whether the sweepnet samples, soil samples, or hand- picked samples are considered, there were no significant differences in the seasonal larval parasitism rates in the three crops in either 1974 or 1975. The same is true for the plots at Galien, where soil samples showed no significant difference in the percentage of pupal cells para- sitized by I, julis (Table 42). Parasitism of CLB Eggs CLB eggs found in the square-foot foliage samples which were col- lected twice-weekly at Gull Lake were kept in the laboratory to assess parasitism by Anaphes flavipes. Only in 1974 were enough eggs found to provide meaningful results, which are summarized in Table 43. In this table the number of "viable eggs" is equal to the number of larvae which hatched plus the number of eggs which produced adult Anaphes. This last value was not known exactly, since all eggs collected from a given field on a given date were placed in one petri dish together, and the parasites were counted in the dish after emergence. Instead, the number of parasitized eggs was estimated by dividing the number of adult 102 Table 42. Percent parasitism of CLB larvae by I, julis at Galien in 1975 as measured by examining pupal cells collected in soil samples (8 ‘xzyd2 samples per plot). Crop Total Cells Parasitized Cells % Parasitism 0 l8 5 27.8 OA 66 20 30.3 OW 62 15 24.2 Table 43. Percent parasitism of CLB eggs by Anaphes flavipes in the three crops at Gull Lake in l974. Numbers in parentheses are the number of "viable eggs" collected in 30 ft2 of foliage (see text). CROP Date 0 0A OW % (n) % (n) % (0) 5/17 0 (O) O (0) O (0) 5/21 0 (10) 0 (38) 0 (3) 5/25 0 (52) 0 (73) 0 (23) 5/29 0 (125) 0.4 (148.6) 0 (20) 6/3 0.4 (72.3) 0 (181) 1.8 (16.3) 6/7 0.8 (74.1) 10.0 (105.6) 0 (12) 6/12 7.4 (36.7) 9.8 (36.6) 29.9 (113) 6/17 55.9 (13.6) 63.6 (11) 100.0 (.6) 6/21 87.7 (16.2) 74.2 (23.3) 100.0 (2.7) 6/25 100.0 (5.5) 100.0 (5.5) 100.0 (4.2) 6/28 100.0 (23.3) 100.0 (7.9) -- O 7/2 100.0 (3.3) 100.0 (.6) 100.0 (3.6) 7/5 100.0 (1.2) -- O -- 0 Overall rate* 13.6 (433.2) 8.4 (631.1) 15.3 (96.7) average parasitism weighted by the number of viable eggs collected on each date 104 Anaphes produced by 3.3, the mean number of parasites per host as re- ported by Anderson and Paschke (1968). The parasite did not begin to make a substantial impact on the CLB egg population until about June 12, at least a week after peak CLB egg densities were recorded. Thus the overall parasitism rate by this species did not exceed roughly 15%. The DA field had much lower egg parasitism rate than the adjacent 0 field. This may be due to the differing CLB egg densities in the fields (p. 86); the peak egg density in 0A was twice as great as in pure oats. These data indicate that Anaphes is effective at locating hosts even at low host densities, but is slow in building up to effective numbers. When the number of parasites is limited, a higher host density thus gives a lower percent parasitism as in the 0A fie1d. CLB Survival If the area under a graphic plot of insect density vs degree-day accumulations is calculated (by numberical integration) and is divided by the mean number of degree-days that an individual insect is exposed to sampling, then an estimate of the total seasonal production of insects per unit of land is obtained (Southwood 1966, p. 279). Table 44 lists the production of CLB eggs, larvae, pupae, and emerging summer adults for each field in 1974 and of larvae, pupae, and adults in each plot in l975. Egg and larval densities were obtained from foliage samples in 1974, while l975 larval densities were estimated from sweepnet catches (repeated from Table 36). In both years pupal cell and emerging summer adult densities were obtained from evaluations of soil samples. From these values of total production of each developmental stage, 105 Table 44. Total production (no/ftz) of CLB eggs, larvae, pupae, and summer adults in the Gull Lake experimental plots in 1974 and 1975. Year Field Crop Eggs Larvae Pupae Adults 1974 9-16 0 18.66 6.42 0.76 0.30 9-17 0A 27.82 10.23 1.13 0.43 9-11 0W 4.98 1.06 0.15 0.04 1975 9-10 0 -- 0.089 .067 .056 0A -- 0.101 .078 .044 OW -- 0.082 .044 .023 8-10 0 -- 0.284 .022 .011 0A -- 0.203 .033 .022 OW -- 0.223 .056 .011 5-55 0 -- 0.108 .078 .056 0A -- 0.053 .044 .022 OW -- 0.068 .022 .011 Indicates not measured 106 the survival of insects from one stage to the next can be calculated. The interpretation of these survival values depends on the validity of several assumptions which must be made when using this method. First, it must be assumed that samples are collected more frequently than the minimum time that insects are exposed to sampling, so that none are un- observed. For example, CLB larvae require 240 oD>48 for development (8-12 days), so weekly samples should be adequate. Second, it must be assumed that all mortality for a given stage occurs at the end of that stage, so that the calculated total production represents the number of insects entering that stage. When this is true, then the ratio of the p0pu1ation estimates for two succeeding stages gives the proportion surviving through the first stage. If, instead, mortality occurs earlier'GOr example, at a constant rate throughout a life stage)then the total production estimate actually represents the number of insects of median age for that stage. A problem then exists in that the calcu- lated total production is an underestimate of the number of median-aged insects since the duration of exposure to sampling is actually less (because of mortality), on the average, than the developmental time used in the calculation. How this error source affects the survival rates which are calculated from the total production is complex and is not well understood at this time. For this discussion it is merely assumed that the survival values represent the proportion of insects surviving from the beginning of one stage to the beginning of the next. This may not be too inaccurate since in insects the time of greatest hazard and mortality is likely to be at ecdysis, which by definition occurs at the end of a life stage. 107 Table 45 gives the survival rates for each life stage, calculated from the seasonal productions reported in Table 44. The interpretation of these survival rates is difficult. The nature of the statistical distribution of the values is unknown due to the complex method of calculation; therefore, classical statistical tests must be applied with caution. An analysis of variance of the larval and pupal survivals (for both years combined) showed no significant differences among the crop treatments. However, in every case pupal survival was lowest in the OW plots (Table 45), perhaps indicating that a real effect exists. No clear-cut differences in soil moisture (p. 55) or temperature (p. 57) in the OW plots during the time of pupation were observed. Recall that parasitism by I, julj§;-included in pupal mortality--was higher (although not significantly) in the ON plots (Tables 40 and 41). Perhaps the parasites were more able to locate hosts in the OW plots where stem density and plant biomass were lower (Table 9). This is an interesting possibility which might be looked at further. The higher larval survivals obtained in 1975 are to be expected, since in that year the larval densities were estimated from sweepnet catches. It is known that the sweepnet fails to capture most of the younger larvae, thus underestimating the density. The interpretation of sweepnet samples is a difficult problem, and has been the subject of other research efforts (Fulton 1975, Logan 1976). The approach of Ruesink and Haynes (1973) (i.e., multiplying larvae per sweep by 1.02 to obtain larvae per ftz) has been used here for simplicity, but it is recognized that the density estimates thus obtained are subject to considerable error whose nature is hard to evaluate. 108 Table 45. Survival of CLB life stages at Gull Lake. Year Field Crop Eggs Larvae Pupae 1974 9-16 0 .344 .118 .395 9-17 0A .368 .110 .381 9-11 ON .213 .142 .267 1975 9-10 0 -- .753 .836 0A -- .772 .564 OW -- .537 .523 8-10 0 -- .077 .500 0A -- .163 .667 OW -- .251 .196 5-55 0 -- .722 .718 0A -- .830 .500 OW -- .324 .500 Overall mean 0 " .418 .612 0A " .469 .528 ON " .314 .372 109 Field Cage Study To obtain higher egg densities for evaluating within-generation survival, in early June, 1975, 240 adult beetles were stocked in each of two milli-acre fie1d cages in each plot of field 9-10. Cages were moved about a week later and eggs were counted in one ft2 sample in each quarter of each cage. The egg counts on June 16 were considered to represent the total egg input to the caged areas since the stocked eggs did not have sufficient time to hatch prior to counting and the endemic p0pulation was extremely small in comparison. In July pupal densities were determined in each cage quarter by taking a 1/2-yd2 soil sample. From these data survival rates from egg to pupa were calculated. The survival percentages for each cage- quarter are plotted in Fig. 17 as a function of the log of initial egg density in that quarter, and the data are reported in Table 46. The several cases of zero survival are probably artifacts of sampling at low pupal densities where at most one or two cells per sample are found. The probability of finding no pupal cells in a sample when the mean density is less than 1.0 is fairly high. Therefore, these points (except for one) were excluded from the regression line in Fig. 17 and from the mean survivals calculated for the CLB in the different crops (Table 46). There is evidence here (Fig. 17) for density-dependent mortality, with percent survival to the pupal stage inversely related to the log of initial egg density. This was also reported by Casagrande (1975) and Helgesen and Haynes (1972) in experiments similar to this one. Helgesen and Haynes suggested that disturbance of the leaf surface 9;, SURVIVAL TO PUPA 110 51 A A O 0 0A 4;. <> ()Vv 31 Vale-3.6211 2* z: r .60 '4 O 5 10 '20 30‘50'1'100 EGGS PER FT2 (log scale) Figure 17. Percent survival to the pupal stage of eggs ovi- posited by adult CLB caged for one week in the three crops at Gull Lake in l975. 111 Table 46. Survival of CLBs from egg to pupa in the field-cage study at Gull Lake in l975. Crop Sample Eggs/ft2 Pupae/ft2 % Survival Mean survival 0 1 14.6 0 -- 2 7.7 0.22 2.86 3 18.0 0.22 1.22 4 7.7 0 -- 5 9.4 0.44 4.68 2.08 6 12.9 0 -- 7 22.3 0.22 0.99 8 35.1 0.22 0.63 0A 1 22.3 0.44 1.97 2 9.4 0.22 2.34 3 10.3 0 -- 4 16.3 0 -- 5 12.9 0 -- 1.65 6 12.9 0 -- 7 23.1 0.22 0.95 8 16.3 0.22 1.35 0W 1 7.7 0.22 2.86 2 40.3 0 0.00 3 57.4 0.22 0.38 4 14.6 0 -- 5 21.4 0.44 2.06 1.44 6 12.9 0.22 1.71 7 14.6 0 -- 8 13.7' 0.22 1.61 112 caused by adult and larval feeding interferes with first instar estab- lishment on the leaf in a density dependent manner. Alternative hypo- theses mignt include some form of density-dependent parasitism (presumably by A, flavipes), predation, or dispersal of larvae. The low survivals reported here, as compared to those found by Casagrande (1975), may be due to a higher rate of egg parasitism. The beetles used in this study oviposited in June, when parasitism by A, flavipes increased rapidly (Table 43), while in Casagrande's study the eggs were laid in May when the wasp was not present. There is no indication that survival to the pupal stage was different in any of the three crops (Fig. 17). Emergence of T. julis Data on the cumulative percent emergence of I, jglj§_adults from the soil in oat fields in 1974 is summarized in Table 47 and is plotted (with least-squares regression lines) on a probit scale in Fig. 18. A test for homogeneity of the slopes of the regression lines showed that they were not significantly different (p>.05). T-tests comparing the data after adjustment for the regression effect (common slope = .0098) to the mean 00 value (1318) (Ostle 1963, p. 199) showed the both OA and OW had significantly higher cumulative percent emergences at 1318 oD>48 than did pure oats. These results indicate that the rate of emergence (% per 00) was similar in all these crops but emergence occurred earlier in the oats planted with companion crops than in pure oats. The 0D accumulations at 50% emergence were 1322, 1269,and 1255 for 0, 0A and OW, respectively. This represents a difference of about three days. A regression approach is not strictly an appropriate analytical 113 .0808 :8 88080 P880828880x8 808 0088 888008 8.83.0 .H .80 8888 08888080083 8808088808 8088880 85883050 .08 8800.88 vaoo 8m. . 00.8. . 000. . 00.8. . 008. 00.. . . _ 8 n \ .8 \\ o .0. «2.083.018» 3011611 5.0.8888...” 6.1-4.--. 6.... .08 .n 8008888...» olol. .K . xv... 08 \s O .08. .00 "mmw .mm .000 .mwm 30N3983W3 °/o BAILV'IleflO 114 Table 47. Cumulative percent emergence of I, julis adults in 10 l yd2 emergence traps placed in each experimental oat field in 1974. 0 CROP Date D>48 0 0A 0w 6/29 1065 0.0% 0.0% 0.0% 7/1 1123 3.3 3.2 11.2 7/2 1157 4.5 6.2 11.2 7/3 1195 6.2 25.3 25.5 7/4 1229 6.6 33.1 41.8 7/5 1249 22.9 55.3 47.6 7/8 1333 68.8 77.2 72.6 7/9 1369 77.8 90.5 83.7 7/10 1403 86.2 92.9 100.0 7/11 1425 88.8 96.7 7/12 1448 96.3 98.7 7/15 1547 98.0 100.0 7/16 1569 98.0 7/17 1595 100.0 Total Caught 87 194 43 115 procedure for cumulative emergence data since successive observations are not really independent. An appropriate non-parametric test which requires no assumptions about the underlying distribution of the ob- servations is the Nilcoxon signed rank test (Ostle 1963, p. 468). This test, too, showed that given emergence percentages occurred earlier in OA and ON than in pure oats (T = 78, T l; n = 12, and p <.01 O-OA o-ow = for both comparisons). The summer emergence (second generation) of I, juli§_is not well synchronized with the occurrence of CLB larvae. Parasite emergence does not begin until most of the CLB larvae have pupated (Fig. 19). Thus, earlier emergence, as in the companion fields, should increase the effectiveness of I, julj§_as a biological control agent. Wasps which would ordinarily die without ovipositing would be able to produce progeny if they emerged early enough to locate hosts. However, as can be seen in Fig. 19, even in the companion crops second generation L. jy_l_i_s emergence was too late to be of any consequence. In 1975 the density of emerging I, julj§_was too low (Table 48) to permit analysis of emergence rates. FeedingiDamage CLB feeding damage to oat leaves, as measured by the number of adult and larval feeding scars per leaf at the end of the larval period, was not related to the leaf position for the top, (flag) second, or third leaves (Table 49A). Therefore, data for all three leaves were combined in comparing the different fields. In the experi- mental fields, the degree of feeding damage was related to the number 116 aowaoaawa % BAILV'lnlNanO snnr 1' .enmp 0w Awumpsszu 000 00>00F m00 00 0000000000 00 A 0. 000. 00..... . 000. 00.0. 00.0 00.0 .x 00. .. a. 0. 00. .. .. 00. 0.... H... 0.0 00. .0. 00.. 1--.... .0. 0000.. 00¢ 31:1 / BVAHV'I 8'10 117 Table 48. Emergence (number and cumulative percent) of second generation I, Julis adults in 30 1 yd2 emergence traps placed in each crop in 1975. CROP Date 00>“ No. 0 % No.0A % “RESET 6/25 985 2 5.9 3 30.0 0 0 6/26 1007 9 32.4 2 50.0 0 0 6/27 1031 13 70.6 2 70.0 1 25.0 6/30 1109 3 79.4 3 100.0 0 25.0 7/1 1134 6 97.1 0 1 50.0 7/2 1159 1 100.0 0 1 75.0 7/3 1186 0 O 1 100.0 Total Caught 34 10 4 118 Table 49A. ANOVA results (mean squares) for damage to the top three leaves of oats in 1974 and 1975. Error mean squares are followed by their degrees of freedom (in parentheses). 1974 1975 Source of Variation df Holes/leaf % Green Holes/leaf Crop (0, 0A, 0W) 2 16267** 7191** 59.5 Leaf Position 2 752 l6046** 57.9 Interaction 4 321 351 12.2 Error 505 (430) 304 (429) 22.7 (261) 119 of larvae present. Therefore, in 1974 0A was most damaged and ow was least damaged, while in 1975 there were no significant differences in either the larval densities or the amount of feeding damage (Table 49B). The percent of the leaf surface area which was estimated to be green in 1974 was significantly affected by both the crop type and the leaf position (Table 49A). Thus, the leaves for 0A, which had the largest larval population, were significantly less green (Table 498). The mean percent green for the top, second, and third leaves, averaged over all three fields, were 92.4%, 93.3%, and 74.9%, respectively. The effect of leaf position is probably due to senescence of the lower leaves. Alfalfa in the Second Year The two methods of alfalfa establishment considered here were: (1) spring seeding with an oat companion crop, and (2) fall seeding alone. In 1974, field 9-17 (at Gull Lake) was planted to oats with inter-seeded alfalfa. In the fall of 1974 the adjacent oat stubble field, 9-16, was plowed and seeded to pure alfalfa. Thus in 1975 the fields were available for comparison of the two establishment methods. Table 50 gives the results of measurements made on alfalfa stem density, stem length, and wet weight of aerial plant parts for both alfalfa and for all other plant species (weeds). Comparisons by t-tests of the means of each variable for the first and last sampling dates gave the following results: On May 9, 1975, there were significantly longer stems and a greater total weight of alfalfa foliage and weight per alfalfa stem in the companion-planted field (p<.01). The stem density was also greater at the 6% level of significance. There was no difference 120 Table 498. Seasonal larval density and subsequent feeding damage to oat leaves in the experimental fields in 1974 and 1975 [Mean number of CLB larval feeding scars per leaf (both years) and estimated percent of leaf surface area which was green (1974 only), averaged over the top three leaves]. 1974 1975 Crop Larvae/ft2 Holes/leaf % Green Larvae/ft2 Holes/leaf O 6.42 16.0 a 89.9 b .09 0.9 a 0A 10.23 27.4 b 79.0 a .10 2.4 a ON 1.06 6.7 c 91.8 b .08 2.1 a Means in the same column and followed by the same letter are not significantly different at the 5% level (Duncan's NMR). 0 u 0 .4. 0.000.0 0.000.00 0. 0000. 00 0.000.00 0.000.00 0000 0.000.0 0.000.00 0.000. 00 0.000.00 0.000.00 0000 0.000.0 0.000.00 m. 000. m0 0.000.00 0.000.00 0000 00.00.0 0.000.00 0.000. 00 0.000.00 0.000.00 0000 00.00.0 0.000.00 0.000.00 0.000.00 0.000.00 0000 00.00.0 0. 00m. 00 0.000.00 0.000.0 0.000.00 000 000000 00000 0000000 H070 1. 0.000.0 m. 0000. 00 0. 0000. 000 0. 000. 00 0.000.00 0000 mm 00.000.0 0. 0000. 000 0. 0000. 000 0.000. 00 0.000.00 0000 0.000.m 5 0000. 000 0. 0005 00 0.000.00 0.000.00 000m «0.000.0 0. 0000. 00 0.000 .00 0.00 .00 0. 005 00 0000 0,000.0 0. 0000. 000 0 0000. 000 0.000.00 0.000. 00 0000 000.000.0 0.000. 00 0. 0000. 00 0.000.00 00.000.00 000 00000000 000000000 000 0003 ”00-0 0000000\000002 00003 000000< 000 000000 000000000 0000 00000.0 000 000003 003 .000 .0.0 0 my 0000000 030 00 0000 0000>000 000 0000 00:0 00 00000000000 000000 0000000 000 00 003000 00000 00 00000000 0000 .00 00000 122 in the weight of weed foliage. On May 29, only the weight per stem of alfalfa was significantly greater in the companion-planted field (total wet weight of alfalfa was also greater, but only at the 13% level of significance). Thus, while the spring-seeded alfalfa got an earlier start on growth in the second year, by the end of May few differences persisted. When the fields were cut on June 30, field 9-l6 yielded 328 bales and field 9-l7 yielded 312 bales. The two fields were the same size and yielded about the same amount of hay. Although the total weight of weed foliage was similar on May 29 in the two alfalfa fields, the species-composition of these weeds was very different (Table 51). Field 9-16, the fall-planted pure stand, had a greater amount of grasses, Lychnis, crucifers, and other weeds. Field 9-l7, which was spring-planted with an oat companion crop, had more clovers and dandelions. The latter situation is probably more desirable for the production of forage, since red and white clover are, themselves, forage crops. This may have resulted from impure alfalfa seed, or, more likely, is the result of natural successional processes. The alfalfa weevil population was extremely low at Gull Lake in 1975; very few eggs, and no pupae, were found while examining the ft2 alfalfa foliage samples (Table 52). While no statistical comparisons were attempted with these sparse data, it appears that the heavier stems of the alfalfa in field 9-l7 (companion planted) harbored more and larger egg masses. 123 Table 51. Total wet weight (g) of aerial portions of weeds in ten 1 ft2 samples in the two alfalfa fields on May 29, 1975. Fall Spring (Without Oats), (With Oats) Weed Class g % g % Grasses 528.9 59.6 ll4.l l5.l Clovers (red and white) 0.8 O.l 566.5 74.7 Lychnis sp 96.4 10.9 50.0 6.6 Cruciferae 164.6 l8.6 2.7 0.4 Taraxacum officinalis 4.9 0.6 l6.0 2.1 Other (Chickweed, Oxalis sp, 9l.2 10.3 8.8 l.2 Viola sp, others . TOTAL 886.8 100.0 758.1 100.0 124 Table 52. Number and size of alfalfa weevil egg masses in ten 1 ft2 foliage samples taken in the two alfalfa fields on several dates in l975. Each entry gives the number of eggs in an egg mass; masses enclosed by parentheses are from the same ft2 sample. Fall Spring Date (Without Oats) (With Oats) 5/9 no eggs 5 5/l3 3 14 5/l6 no eggs l4,(4,7),7,(ll,6,8) 5/20 6,(l,6) (6,2) 5/23 4 (12.9) SUMMARY AND CONCLUSIONS This study has been an attempt to examine, in some detail, the im- pact of a companion cropping scheme on the insect and plant components of a pest/crop system. In Michigan it is an equally common practice to interplant oats and alfalfa as it is to establish oats alone. However, to my knowledge all of the research to date on the cereal leaf beetle has been done on pure stands of oats. The potential influence of a leguminous companion crop on the distribution, production, and survival of the CLB remained unevaluated prior to this present study. Unfortunately the complicating factors of very low CLB densities at the project location and large year to year and field to field variations in climatological and other conditions contributed to results which are confusing and difficult to evaluate. The procumbent growth of non-vernalized wheat did not seem partic- ularly attractive to the CLB and so did not act as a trap crop under the observed conditions. Under a heavier infestation, this interplanted non-crop may serve as an egg sink, diverting enough early ovipostional activity from the oat crop to prevent severe losses to the pest. The growth pattern of oats was statistically different for the three planting combinations (0, 0A, 0W), but the apparent influence of the companion crop varied greatly both within and between years. At the end of the season in l974, OW (compared to pure oats) had a higher 2 and stem density, greater height, greater foliage wet weight per ft per stem, greater leaf surface area, and a 52% greater yield of grain. 0A produced a greater weight per stem and leaf surface area, and a 4l% 125 126 higher yield. In l975, 0W stems were fewer and shorter, and the plants had a lower weight per ft2 and per stem, leaf surface area was greater, and yield was equal to or less than that of pure oats. 0A plants again had a greater weight per stem and leaf surface area, but yield was not different. Since, in l975, the effect of differences between fields was found to be greater than the treatment effects, caution should be exercised in interpreting the l974 results where each treatment was assigned to a separate field. Despite this note of caution, I feel that the enhanced growth observed in the companion planted fields in l974 was in part a real treatment effect. It is concluded that clima- tological, varietal, or soil differences strongly interact with and modify the companion crop effects, much as the outcome of interspecific plant competition is determined by such factors. The effect of the interplanted alfalfa on the total nitrogen con- tent of above-ground plant material was not as great as expected: in only one field did OA plants contain more nitrogen than with oats alone. This probably resulted from an excess of nitrogen (from fertilizer) being present already, with plant growth being limited by some other factor. While the portion of the total nitrogen found in each category of plants (oats, grasses, weeds) and the percent nitrogen in oat foliage did not differ among treatments, at harvest the oat heads in the pure stand did contain slightly more nitrogen. A more meaningful evaluation of the benefits of a leguminous companion crop must include plots not fertilized with nitrogen. Damage (feeding holes) to oats was directly related to CLB densities in all plots, but was never severe enough to noticeably affect yield. 127 CLB egg and larval densities were greater in OA and less in OW, compared to O, in l974. In 1975 there were no differences in egg densities but the larval density in OW was less. In l974 pupal cell density was lower in OW than in the other two crops, and the density of emerging adults was higher in 0A. In 1975 there were no differences for these stages. Results from plots at Galien and Collins Rd. pre- sented still other outcomes with no clear pattern to the overall results except perhaps a lower seasonal production of CLBs in the OW plots. This combination may not be as attractive for some reason, but no definite conclusions can be drawn. There was no effect of the type of planting arrangement on parasitism of CLB larvae by Tetrastichus julis. A lower rate of egg parasitism by Anaphes flavipes in 0A was observed in 1974, but I believe this may simply be due to the higher egg density in that plot early in the season when parasite numbers were limited. There is some evidence that emergence of second generation I; julis adults occurs earlier in the companion-planted fields, but the effect is probably insufficient to improve the synchrony of the parasite enough to increase its impact on the CLB population or to significantly increase the number of overwintering wasps. Conditions of soil moisture and temperature in 1975 did not differ among treatments in any way that might suggest a cause for earlier emergence from the soil in the companion plots. Under natural field conditions and with higher egg densities obtained by caging ovipositing beetles, no statistical differences were found of survival rates of eggs, larvae or pupae in the 128 experimental plots (there was a slight trend of lower pupal survival in the OW plots). With alfalfa established by the two common methods which were used in this study, there were essentially no differences in hay production by the time of first harvest. Spring establishment with an oat companion resulted in forage with a greater proportion of red and white clover, as opposed to grasses and crucifers, in the weed component, and this may be desirable. While alfalfa weevil numbers were too low to provide much data, it appears that the companion-planted alfalfa was more attractive to the weevil in the second year. In conclusion, under the conditions of low CLB densities observed in this study, companion planting of oats gives rise to only minor differences, if any, in the numbers, survival, and feeding damage of cereal leaf beetles. The effect on the growth and yield of oats is quite variable and is probably small compared to the effects of weather and field conditions and plant variety. The use of an alfalfa companion may supplant the application of nitrogenous fertilizer, but this was not evaluated. Models recently developed to simulate the cereal leaf beetle/ small grain system for pest management purposes (eg. Tummala et al l975, Jackman 1976) should be equally valid whether oats are planted in pure culture or with a companion crop. LITERATURE CITED Agboola, A. A. and A. A. Fayemi. 1970. Interplanting of maize with legume. II. The effect of phosphorus and intercropping of tropical legumes on the yield of maiZe. W. Afr. J. Biol. Appl. Chem. No. 2. Anderson, R. C. and J. D. Paschke. 1968. The biology and ecology of Anaphes flavipes (Hymenoptera: Mymaridae), an exotic egg parasite of the cereal leaf beetle. Ann. Entomol. Soc. Amer. 61: 1-5. Anonymous. 1960. Root exudates. Commonwealth Sci. & Indust. Res. Org. Australia. No. 33: 24- 7. Arny, A. C., T. E. Stoa, C. McKee, and A. C. Dillman. l929. Flax crop- ping in mixture with wheat, oats, and barley. USDA Tech. Bull. 133. 48 p. Atsatt, P. R. and D. J. O' Dowd. 1976. Plant defense guilds. Science 193: 24-29. Beirne, B. P. 1967. Pest management. Leonard Hill, London. 123 p. Bliss, C. I. and R. A. Fisher. 1953. Fitting the negative binomial distribution to biological data, and note on the efficient fitting of the negative binomial. Biometrics. June: 176-200. Casagrande, R. A. 1975. Behavior and survival of the adult cereal leaf beetle, Oulema melanopus (L.). PhD Thesis, Michigan State Univ., East Lansing, Michigan. 174 p. Chestnutt, D. M. B. 1972. The effects of white clover and applied nitrogen on the nitrogen content of various grass/clover mixtures. J. Br. Grasslnd. Soc. 27(4): 211-6 Christozov, At. 1968. Study on raising plant protein production by growing maize in mixed crops with soybeans and broad beans. Rastenievadni Nauki. 5(1): 91-104. (Russian, Engl. Summary). Cowling, D. W. and D. R. Lockyer. 1967. A comparison of the reaction of different grass species to fertilizer nitrogen and to growth in association with white clover. 11. Yield of nitrogen. J. Br. Grasslnd. Soc. 22: 53—61 DeLoach, C. J. and J. C. Peters. 1972. Effect of strip-planting vs solid-planting on predators of cotton insects in southeastern Missouri, 1969. Environ. Entomol. 1(1): 94-102. Denisen, E. L. 1958. Principles of horticulture. MacMillan Co. 129 130 Dijkstra, J. and A. L. F. deVos. 1972. The evaluation of selections of white clover (Trifolium repens L.) in monoculture and in mixture with grass. Euphytica. 21(3): 432-49. Eisenhart, C. 1947. The assumptions underlying the analysis of variance. Biometrics. 3: 1-21. Evans, M. W. 1916. Some effects of legumes on associated non-legumes. J. Am. Soc. Agron. 8(6): 348-57. Flanagan, T. R. and J. B. Washko. 1950. Spring grain characteristics which influence their value as companion crops. Agron. J. 42: 460. Frame, J., R. D. Harkess, and I. V. Hunt. 1972. The effect of a rye- grass companion grass and the variety of red clover on the pro- ductivity of red-clover swards. J. Br. Grasslnd. Soc. 27(4): 241-9. Franz, J. 1961. Biological control of pest insects in Europe. Ann. Rev. Entomol. 6: 183-200. Fulton, W. C. 1975. Monitoring cereal leaf beetle larval populations. M.S. Thesis, Michigan State Univ., East Lansing, Michigan. 108 p. Gage, S. H. 1972. The cereal leaf beetle, Oulema melanopus (L.), and its interaction with two primary hosts: winter wheat and spring oats. M.S. Thesis, Michigan State Univ., East Lansing, Michigan. 105 p. Gage, S. H. 1974. Ecological investigations on the cereal leaf beetle, Oulema melanopus (L.), and the principal larval parasite, Tetrastichus julis (Walker). PhD Thesis, Michigan State Univ., East Lansing, Michigan. 172 p. Gage, S. H. and D. L. Haynes. 1975. Emergence under natural and manip- ulated conditions of Tetrastichus julis, an introduced larval parasite of the cereal leaf beetle, with reference to regional population management. Environ. Entomol. 4(3): 425-34. Gillespie, J. 1964. Peacock manure and marigolds. The Vicking Press, N.Y. Gouder, B. A. Y. and S. V. Patil. 1971. Possibilities of growing mixed crOps of wheat with cotton. Mysore J. Agric. Sci. 5(4): 496-9. Grodzinskij, A. M. 1969. On metabolite exchange between neighboring plants. Qual. Plant Mat. Veg. 17(2): 93-100. (German, Engl. summary). Haggar, R. J. 1970. Use of companion crops in grassland establishment in Nigeria.. Samaru Res. Bull. 108: 47-52. Haynes, D.L. 1973. Population management of the cereal leaf beetle. Insects: Studies in population management. Ecol. Soc. of Australia, Memoirs I, Canberra. 131 Helgesen, R. G. and D. L. Haynes. 1972. Population dynamics of the cereal leaf beetle, Oulema melanopus, (Coleoptera: Chrysomelidae): A model for age specific mortality. Can. Entomol. 104: 797-814. Hildebrand, S. C. and L. 0. Copeland. 1975. Seeding practices for Michigan crops. Coop. Ext. Bull. E-489, Michigan State University. 5 p. Hoxie, R. P. and S. G. Wellso. l974. Cereal leaf beetle instars and sex, defined by larval head capsule widths. Ann. Entomol. Soc. Amer. 67(2): 183-6. Humm, D. J., R. S. Fulkerson, and W. E. Tossell. 1969. Seedling year management of alfalfa-grass mixtures established without a compan— ion crop. Can. J. Plant Sci. 49: 477-81. Hunter, 8. T. 1971. Gardening without poisons. Houghton Mifflin & Co., Boston. 318 p. J. C. 1973. The technique that halves your insect population. Organic Gardening & Farming. May: 103-4. Jackman, J. A. 1976. A quantitative description of oat growth with cereal leaf beetle populations. PhD Thesis, Michigan State Univ., East Lansing, Michigan. 172 p. Johns, Glinn F. (ed.) 1966. The organic way to plant protection. The staff of Organic Gardening and Farming magazine. Rodale Press Inc. Emmaus, Pa. Kappel, B. 1973. Marigold magic works in Montana. Organic Gardening & Farming. May: 92. Kilcher, M. R. and D. H. Heinrichs. 1960. The use of cereal grains as companion crops in dryland forage crop establishment. Can. J. Plant Sci. 40: 81-93. Kust, Cyril A. 1968. Herbicides or oat companion crop for alfalfa establishment and forage yields. Agronomy J. 60(2): 151-4. Lawrence, T. 1967. Effect of a wheat companion crop on the seed yield of Russian wild rye grass. Can. J. Plant Sci. 47(5): 585-92. Lawrence, T. 1970. Effect of a wheat companion crop on the seed and dry matter yield of crested wheat grass. Can. J. Plant Sci. 50(1): 81-6. Loehwing, W. F. 1937. Root interactions of plants. Bot. Rev. 3: 195-239. Logan, P. 1976. (untitled). PhD Thesis, Michigan State Univ., East Lansing, Michigan. (in preparation). 132 Lyon, T. L. and J. A. Bizzell. 1913a. A discussion of certain methods used in the study of “the associative growth of legumes and non- legumes." J. Am. Soc. Agron. 5(2): 65-72. Lyon, T. L. and J. A. Bizzell. 1913b. Is there a mutual stimulation of plants through root influence? J. Am. Soc. Agron. 5(1): 38-44. Margalef, Ramon. 1968. Perspectives in ecological theory. U. of Chicago Press. 111 p. McClelland, C. K. 1928. The effect of interplanted legumes on the yields of corn. Ark. Agr. Exp. Sta. Bull. No. 229. McGowan, A. A. and W. A. Williams. 1971. Growth of subterranean clover established with a cereal companion crop. Agron. J. 63(4): 643-6. Mendenhall, W. 1968. Introduction to linear models and the design and analysis of experiments. Duxbury Press. 465 p. Milthorpe, F. L. and J. Moorby. 1974. An introduction to Crop Physiology. Cambridge U. Press, London. 202 p. Morris, R. F. 1955. The development of sampling techniques for forest insect defoliaters, with particular reference to the spruce bud- worm. Can. J. Zool. 33(4): 225-94. Morrish, R. H. 1934. Crop mixture trials in Michigan. Mi. Agr. Exp. Sta. Special Bull. No. 256. 11 p. Murdoch, W. W., F. C. Evans, and C. H. Peterson. 1972. Diversity and pattern in plants and insects. Ecology. 53(5): 819-29. Neter, J. and W. Wasserman. 1974. Applied linear statistical models. Regression, analysis of variance, and experimental designs. Richard D. Irwin, Inc., Homewood, Illinois. 842 p. Odum, E. P. 1971. Fundamentals of ecology. 3rd ed. W. B. Saunders Co. 574 p. Olds, Jerome (compilation supervisor). 1959. The encyclopedia of organic gardening. The staff of Organic Gardening and Farming magazine. Rodale Books, Inc. Emmaus, Pa. Olsen, F. J. and E. R. Tiharuhondi. 1972. The productivity and botan- ical composition of some selected grass/legume pasture mixtures at different seeding rates. E. Afr. Agric. For. J. 38(1): 16-22. Ontario Ministry of Agriculture and Food. 1971. Agricultural statis— tics for Ontario. Publication 20. 111 p. Ostle, B. 1963. Statistics in research. Iowa State University Press. 585 p. 133 Perkins, S. 0. and J. Tyson. 1926. Soil survey of Kalamazoo County Michigan. USDA Bureau of soils. U.S.G.P.O., Washington, D.C. Peters, R. A. 1961. Legume establishment as related to the presence or absence of an oat companion crop. Agron. J. 53: 195-8. Petrakieva, I. and T. Naidenov. l968. Yields and feeding value of annuals grown alone and as mixed crops for silage. Rastenievadni Nauki. 5(3): 78-91. (Russian, Engl. summary). Petrov, G. 1968. On mixed crops of barley and wheat. Rastenievadni Nauki. 5(7): 23-32. (Russian, Engl. summary). Philbrick, H. L. and R. B. Gregg. 1967. Companion plants and how to use them. Devin-Adair. Old Greenwich, Conn. 113 p. Philbrick, J. H. and H. L. Philbrick. 1963. Gardening for health: an introduction to the method of bio-dynamic gardening. St. George Books, Blauvelt, N.Y. 93 p. Pielou, E. C. 1974. Population and community ecology: principles and methods. Gordon and Breach Science Publishers. 424 p. Pimentel, D. 1961. Species diversity and insect population outbreaks. Ann. Entomol. Soc. Amer. 54(1): 76-86. Pimentel, D. 1970. Population control in crop systems: monocultures and plant Spatial patterns. Proc. Tall. Timb. Conf. on Ecol. Anim. Contr. by Hab. Mgmt. Feb. 26-28, 1970: 209-21. Poston, F. L. and L. P. Pedigo. 1975. Migration of plant bugs and the potato leafhopper in a soybean-alfalfa complex. Environ. Entomol. 4 l : 8-10. Robinson, R. R., J. H. Young, and R. D. Morrison. 1972a. Strip-cropping effects on abundance of predatory and harmful cotton insects in Oklahoma. Environ. Entomol. 1(2): 145-49. Robinson, R. R., J. H. Young, and R. D. Morrison. 1972b. Strip-cropping effects on abundance of Heliothis-damaged cotton squares, boll placement, total bolls, and yields in Oklahoma. Environ. Entomol. 1(2): 140-5. Robinson, R. R., J. H. Young, and R. D. Morrison. 1972c. Effects of Heliothis spp. and strip cropping on cotton lint quantity and quality in Oklahoma. Environ. Entomol. 1(1): 109-ll. Rudd, Robert L. 1964. Pesticides and the living landscape. The U. of Wisconsin Press, Madison, Wisconsin. Ruesink, W. G. and D. L. Haynes. 1973. Sweepnet sampling for the cereal leaf beetle, Oulema melanopus. Environ. Entomol. 2(2): 161-72. 134 Salisbury, F. B. and C. Ross. 1969. Plant physiology. Wadsworth Publishing Co., Inc. 747 p. Schlinger, E. I. and E. J. Dietrick. 1960. Biological control of insect pests aided by strip-farming alfalfa. Calif. Agr. 14(1): 8‘99 15. Scholl, J. M. and J. T. Medler. 1947. Trap strips to control insects affecting alfalfa seed production. J. Econ. Entomol. 40: 448-50. Sevacherian, V. and V. M. Stern. 1974. Host plant preferences of lygus bugs in alfalfa-interplanted cotton fields. Environ. Entomol. 3(5): 761-6. Shah, V. H. 1965. Influence of companion crops on the yield and quality of Berseem (Trifolium alexandrinum Juslen). Indian J. Agric. Sci. 35(4): 291-304. Sokal, R. R. and Rohlf, J. 1969. Biometry. W. H. Freeman and Co. 776 p. Son, S. H. 1969. Effects of intercropping of Sorgo (Andropogon sorghum) and soybean (Glycine max) on growths, yields, and quality of the two crops. Res. Rep. Office Rural Develop, Korea. 12(1): 117-23. Southwood, T. R. E. 1966. Ecological methods. Methuen and Co., London. 391 p. Stanton, T. R. 1929. Oats in the north central states. USDA Farm Bull. No. 1581. 27 p. Steel, Robert G. D. and James H. Torrie. 1960. Principles and pro- cedures of statistics. McGraw Hill. 481 p. Stern, V. M. 1969. Interplanting alfalfa in cotton to control lygus bugs and other insect pests. Proc. Tall Timbers Conf. Ecol. Anim. Contr. Hab. Mgmt. 1:55-69. Stride, G. O. 1969. Investigations into the use of a trap crop to protect cotton from attack by Lygus vosseleri. J. Entomol. Soc. S. Afr. 32(2): 469-77. Syme, J. R. and P. M. Bremner. 1969. Growth and yield of pure and mixed crops of oats and barley. J. Appl. Ecol. 5(3): 659-74. Taras, M. J., A. E. Greenberg, R. D. Hoak, and M. C. Rand (eds). 1971. Standard methods for the examination of water and waste water. 13th ed. Am. Public Health Assoc., Washington D.C. 874 p. 135 Tummala, R. L., W. G. Ruesink, and D. L. Haynes. 1975. A discrete component approach to the management of the cereal leaf beetle ecosystem. Environ. Entomol. 4:175-86. Tyker, H. 1970. Organic gardening without poisons. Van Nostrand Reinhold Co., N.Y. Wellso, S. G. and C. E. Cress. 1973. Intraspecific competition of the cereal leaf beetle reduced through spatial separation of eggs and feeding. Environ. Entomol. 2:791-2. Westgate, J. M. and R. A. Oakley. 1914. Percentage of protein in non-legumes and legumes when grown alone and in association in field mixtures. J. Amer. Soc. Agron. 6:210-5. Young, D. J. 1973. Companion planting beats the bugs--plus. Organic Gardening and Farming. August: 67-8. Zavitz, C. A. 1927. Forty years' experiments with grain crops. Ont. Dept. Agr. Bull. No. 332. APPENDIX A GULL LAKE DEGREE-DAY ACCUMULATIONS 136 137 Table A1. Degree-day accumulations at Gull Lake for 1974. °o>42 Day April May June July 1 O 170 638 1393 2 O 183 657 1433 3 O 197 677 1477 4 O 206 706 1517 5 O 214 736 1543 6 O 222 766 1572 7 O 266 797 1606 8 O 231 833 1645 9 O 233 865 1687 10 O 240 893 1727 11 O 252 907 1755 12 O 262 923 1783 13 O 265 943 1822 14 O 279 967 1867 15 O 296 993 1899 16 O 312 1011 1927 17 O 326 1023 1959 18 O 350 1046 1999 19 O 369 1075 2039 20 12 389 1103 2069 21 34 415 1137 2094 22 50 445 1165 2126 23 53 468 1182 2151 24 58 486 1200 2183 25 64 500 1220 2213 26 78 512 1245 2249 27 100 525 1270 2287 28 122 543 1296 2320 29 142 569 1324 2320 30 158 596 1360 2320 31 620 2320 Table A1 (cont'd). 138 00 > 48 Day April May June July 1 0 114 428 1003 2 0 123 441 1037 3 0 132 455 1075 4 0 137 478 1109 5 O 142 502 1129 6 0 146 526 1152 7 0 147 551 1180 8 0 149 581 1213 9 0 149 607 1249 10 0 153 629 1283 11 0 160 637 1305 12 0 165 647 1327 13 0 165 661 1360 14 0 174 679 1399 15 0 185 699 1425 16 O 195 711 1447 17 0 204 717 1473 18 0 222 734 1507 19 0 235 757 1541 20 8 249 779 1565 21 24 269 807 1584 22 34 293 829 1610 23 35 310 840 1629 24 38 322 852 1655 25 41 330 866 1679 26 51 337 885 1709 27 67 345 904 1741 28 83 357 924 1768 29 97 377 946 1768 30 107 398 976 1768 31 416 1768 139 Table A2. Degree-day accumulations at Gull Lake for 1975. °D>42 Day March April May June July 1 0 41 193 810 1592 2 0 41 207 824 1623 3 0 41 224 840 1656 4 0 41 234 860 1684 5 0 41 245 886 1712 6 0 41 261 910 1742 7 0 41 272 922 1772 8 0 41 286 935 1805 9 0 41 303 951 1829 10 0 42 321 971 1855 11 0 43 337 995 1872 12 0 43 357 1017 1891 13 0 46 370 1043 1909 14 0 49 386 1069 1927 15 0 52 400 1096 1953 16 0 57 410 1116 1983 17 4 65 427 1145 2014 18 8 81 448 1177 2046 19 14 88 478 1213 2077 20 19 90 510 1247 2107 21 24 92 540 1279 2135 22 26 100 564 1315 2163 23 29 111 590 1351 2195 24 35 123 620 1383 2227 25 36 129 652 1407 2253 26 36 135 682 1435 2275 27 36 145 704 1465 2303 28 36 147 722 1499 2331 29 36 161 748 1533 2361 30 36 181 774 1561 2395 31 38 794 2431 Table A2 (cont'd). 140 0D > 48 Day March April May June July 1 0 15 91 529 1134 2 0 15 99 538 1159 3 0 15 110 548 1186 4 0 15 114 562 1208 5 0 15 120 582 1230 6 0 15 130 600 1254 7 0 15 137 607 1278 8 0 15 145 615 1305 9 0 15 156 625 1323 10 0 15 168 639 1343 11 0 15 179 657 1354 12 0 15 193 673 1367 13 0 16 201 693 1379 14 0 17 211 713 1391 15 0 18 219 734 1411 16 0 20 225 743 1435 17 2 25 236 771 1460 18 3 35 251 797 1486 19 5 38 275 827 1511 20 7 38 301 855 1535 21 9 38 325 881 1557 22 9 42 343 911 1579 23 10 47 363 941 1605 24 13 53 387 967 1631 25 13 55 413 985 1651 26 13 58 437 1007 1667 27 13 62 453 1031 1689 28 13 62 465 1059 1711 29 13 71 485 1087 1735 30 13 85 505 1109 1763 31 14 519 1793 APPENDIX B RECOVERY EFFICIENCY OF CLB PUPAL CELLS 141 RECOVERY EFFICIENCY OF CLB PUPAL CELLS In population studies of the cereal leaf beetle, estimates of pupal cell density and fate are often desired. Gage (1974) described a method of washing 36 x 18 x 3 in. soil samples to quantitatively assess cell density and mortality factors. The present study was undertaken to determine how efficient the washing procedure is in recovering CLB pupal cells in samples already dug. No attempt was made to analyze the digging efficiency. METHODS Pupal cells obtained the previous year (1974) by rearing CLB larvae over a vermiculite substrate were sorted into three classes: (1) whole cells (no emergence), (2) cells from which CLB adults had emerged (large hole), and (3) cells from which Tetrastichus julis adults had emerged (small hole). This was done to determine if the condition of the cell influences the recovery probability. 150 cells of each class, divided into 3 replicates of 50 cells, were mixed with field soil in the 1/8 in. mesh screened washing boxes (with the exception of one sample which consisted of only 41 cells). The samples were processed as described by Gage. RESULTS Of the initial 441 cells, 269 (61%) were recovered (Table B1). An analysis of variance (one way with 3 replicates) was performed on the data from the three classes of cells. Prior to analysis the percent 142 143 Table 81. Recovery of CLB pupal cells. Rep. Class Initial Recovered %/100 s1n‘1fT' ‘T 1 Whole 50 25 .50 0.785 2 Whole 50 34 .68 0.970 0.57 3 Whole 50 26 .52 0.805 1 CLB 50 30 .60 0.886 2 CLB 50 22 .44 0.725 0.57 3 CLB 50 34 .68 0.970 1 T. julis ' 50 32 .64 0.927 2 T. julis 50 35 .70 0.991 0.70 3 T. julis 41 31 .76 1.059 Overall recovery (n = 441) 0.61 144 recoveries were transformed using an arcsine transformation: Y* (radians) = sin"1 V Y as recommended by Mendenhall (1968: 208) to stabilize the variance. No difference was found in the recovery efficiencies for the three classes of cells (Table 82). CONCLUSION A sizable portion of the CLB pupal cells initially in a soil sample are not recovered after the washing procedure. Whether these cells are lost by passage, whole or broken, through the screens, or are merely not found in the inspection of the washing residue is not apparent at this time. In making estimates of CLB parasitism and other mortality, no bias should result from the washing process since all types of cells are recovered with equal efficiency. 145 Table 82. Analysis of variancea. Source DF SS MS F Treatments 2 .36794E—01 .18397E-ol 1.830 Error 6 .60332E-01 .10055E-01 Total 8 .97126E-01 a Arcsine transformation on percent recoveries. b Not significant. HICHIGQN STRTE UNIV. LIBRQRIES 3110350