0515‘“ Wu ABSTRACT BEHAVIOR AND SURVIVAL OF THE ADULT CEREAL LEAF BEETLE, OULEMA MELANOPUS (L.) By Richard Alfred Casagrande Adults of the cereal leaf beetle (CLB) Oulema melanopus (L.) survive the winter in highest densities at the edge of woodlots. Other habitats include sparse woods, fence rows, dense woods, and cr0plands (in order of "preference"). Beetles overwintering at the ground surface generally do not receive lethal low temperature exposures. However the small proportion overwintering above ground does receive chronic exposures to temperatures below 25°F, the upper threshold for cold-induced mortality. A predictive model is developed to relate mortality to the duration and severity of continuous cold exposures. Experiments also reveal the significance of preconditioning and recovery from successive cold exposures in determining the mortality from low temperatures. Two control features are evaluated which are based on adult cereal leaf beetle behavior. Oviposition is greatly reduced by resistant wheat of the variety Vel. Plantings of mixtures of resistant and susceptible wheat resulted in intermediate reductions in oviposition. Within genera- tion of the beetle seemed unaffected by the field plantings of resistant wheat as did the behavior and survival of the larval parasite Tetrastichus julis (Walker). Richard Alfred Casagrande Strip spraying utilizes the CLB behavioral trait of taking frequent short flights within grain fields in a control program that entails spraying a field with narrow bands of a persistent insecticide. As a result of their normal movement beetles contact the insecticide. A mathematical model is developed which is based on a l-dimension diffusion equation and functions for insecticide decay and effect on beetles. This model simulates various strip spray experiments and shows strip spraying to be an efficient and economical means of reducing adult densities--a conclusion that is verified by field experimentation. Cereal leaf beetle behavior is considered to be a factor in natural population regulation as a shift of the majority of the CLB population from oats to wheat is observed during a phase of declining regional densities. BEHAVIOR AND SURVIVAL OF THE ADULT CEREAL LEAF BEETLE, OULEMA MELANOPUS (L.) By Richard Alfred Casagrande A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Entomology I975 ACKNOWLEDGEMENTS Dr. Dean L. Haynes has been a great major professor and a good friend to me. I shall long be indebted to him for his fruitful imagination, his endless hours of discussions, and his genuine concern for the welfare of his students. I am grateful to Dr. James E. Bath who as department chairman has provided virtually unlimited research opportunities for graduate students. The support I have received from the Department of Entomology in terms of facilities, finances and guidance is most appreciated. To my committee members, Drs. Erik Goodman, Roger Hoopingarner, R. L. Tummala, and James Webster, I am grateful for assistance in organ- izing a meaningful graduate program, for assisting in many aspects of my research, and for providing editorial assistance with this dissertation. Many other people have helped me in many ways. I am especially grateful to Dr. Anihl Kharkar and Mr. David Cobb and fellow students of Dean Haynes, William Ruesink, Stuart Gage, John Jackman, Winston Fulton, and Patrick Logan, whose assistance has influenced much of my work. ii THE BEHAVIOR AND SURVIVAL OF ADULT CEREAL LEAF BEETLES Page LIST OF TABLES ........................... vi LIST OF FIGURES ........................... x PREFACE ............................... l INTRODUCTION ............................ 3 METHODS ............................... 6 The Study Area ......................... 6 Overwintering Sites ...................... 6 Cage Studies of Adult Mortality ................ 7 Population Survey ....................... 9 RESULTS ............................... ll Direct Observation of Overwintering Sites ........... ll Gin Mill Samples ........................ ll Emergence Traps ........................ 13 Distribution in Overwintering Sites .............. 17 Rate of Emergence from Overwintering .............. l7 Regional Emergence at Gull Lake ................ 20 Cage Studies of Adult Mortality ................ 24 Population Survey ....................... 29 Adult Mortality from Survey Data ................ 34 Survey Results versus Overwintering Emergence ......... 37 Cereal Leaf Beetle Behavior .................. 39 Effect of Crop Age on CLB Density ............... 4l DISCUSSION ............................. 42 Overwintering ......................... 42 Density, Behavior and Mortality ................ 44 Implications for Cereal Leaf Beetle Management ......... 50 REFERENCES CITED .......................... 54 APPENDIX A - A Computer Analysis of Strip Spraying for the Control of Cereal Leaf Beetles INTRODUCTION . . .......................... 58 Page METHODS AND RESULTS ......................... 6O Beetle Movement ........................ 60 Theoretical Background and Analysis .............. 6l Insecticide Features: Decay Rate and Insecticide-Induced Mortality .......................... 64 Insecticide Features: Effects on Behavior ........... 73 Temporal and Regional Changes In Insecticide Tolerance ..... 78 Recovery from Insecticide Exposure ............... 80 Strip Spray Model ....................... 86 Field Validation ........................ 87 Model Analysis: Sensitivity .................. 90 Economic Considerations .................... 93 DISCUSSION ............................. 98 APPENDIX B - A Predictive Model for Cereal Leaf Beetle Mortality From Sub-Freezing Temperatures INTRODUCTION ............................ 99 Preconditioning ........................ 99 Regional Differences in Supercooling Ability .......... lOl Predictive Models for Cold-Induced Mortality .......... 102 METHODS ............................... l04 Temperature Control ...................... l04 Cereal Leaf Beetle Treatment .................. l06 RESULTS ............................... 106 Mortality from Constant Exposures - Preliminary Model Development ......................... 106 Seasonal Change in Cold Tolerance ............... ll6 Preconditioning ........................ ll8 Recovery ............................ l24 DISCUSSION ............................. l34 APPENDIX C - The Impact of Resistant Wheat on Populations of the Cereal Leaf Beetle and Its Parasites INTRODUCTION ............................ l38 METHODS ............................... l4O Plot Description ........................ l4O Survey ............................. T42 Cereal Leaf Beetle Behavior .................. l42 Caged Studies ......................... l42 iv Page RESULTS ............................... 145 Survey ............................. 145 Cereal Leaf Beetle Behavior .................. 153 Caged Studies ......................... 155 Distribution Within Cages ................. 155 Within Generation Survival ................ 155 Tetrastichus julis .................... 159 Other Parasites ...................... 159 DISCUSSION ............................. 161 APPENDIX D - The Strip Spray Computer Simulation Used in Appendix A ........................... 164 APPENDIX E - Study Area Surrounding the Kellogg Gull Lake Research Farm, Kalamazoo County, Ross Township ............. 170 Table 10. 11. 12. 13. LIST OF TABLES Cereal leaf beetle adults found in g—square yard samples processed in the gin trash mill in late summer and fall of 1969 .............................. Cereal leaf beetle adults caught emerging from overwintering sites at Gull Lake between 1971 and 1973 using l-square yard emergence traps ........................ Cereal leaf beetle adults caught emerging from overwintering sites near Galien, Michigan, during 1974 and 1975 using 1- square yard emergence traps .................. Densities of cereal leaf beetles caught emerging from over- wintering sites at Gull Lake and Galien, Michigan, in each of the l-square yard emergence traps .............. Analysis of 5 years of individual emergence trap catches com- paring observed frequencies to those expected from a poisson distribution .......................... Cumulative emergence of cereal leaf beetle adults from over- wintering sites as measured by l-square yard emergence cages . . . Determining regional spring emergence of cereal leaf beetle adults ............................. Adult mortality as computed from the 1970 cage study .‘ ..... Adult mortality as computed from cage studies in 1972-1974 . . . . Comparisons of caged adult mortality rates on different crops in 1973 ......................... Seasonal density of adult cereal leaf beetles measured by a sweepnet survey in 1971 and 1972 ............... Seasonal density of adult cereal leaf beetles as measured by a sweepnet survey in 1973 and 1974 ............. Spring adult mortality as computed from the 1971 regional population survey ....................... vi Page 12 14 16 18 19 21 25 26 28 31 Table Page 14. A comparison of peak regional densities of cereal leaf beetle adults in spring and winter grains at Gull Lake from 1971 to 1974 ........................ 40 Al. Observations on cereal leaf beetle movement, 1972—1974, in oats (O) and wheat (W) at Gull Lake .............. 63 A2. Results of two tests on exposing adult cereal leaf beetles to malathion treated plants ................... 65 A3. Minutes exposure required for various mortality levels at different times after insecticide application .......... 68 A4. Regression equations fit to the results in Table A3 ....... 71 A5. Distribution of 20 cereal leaf beetles in a cage at the time of first knockdown ..................... 76 A6. The number of cereal leaf beetles moving out of 20 during a 10 second observation period at various times after intro- duction into cages of sprayed and unsprayed plants ........ 77 A7. Temporal and regional differences in response to topical applications of malathion .................... 79 A8. Live weights of cereal leaf beetles collected in May of 1974 at Galien and Laingsburg .................. 81 A9. Additivity of sequential malathion applications ......... 84 A10. Results of a field test of strip spraying ............ 89 All. An evaluation of 3 strategies which involve spraying of 25% of a field .......................... 96 A12. An evaluation of 5 strip configurations in which 20 ft. widths are sprayed ........................ 97 Bl. Mortality levels resulting from exposures of various tem- peratures and durations ..................... 107 82. Hours exposure required for various mortality levels at various temperatures ....................... 110 B3. Exposure levels corresponding to the time-temperature treatments in Table 82 ...................... 114 B4. Mortality levels resulting from cold exposures at differ- ent times of the year ...................... 117 vii Table BS. B6. B7. BB. 89. BIO. B11. B12. C1. C2. C3. C4. C5. C6. C7. Mortality levels resulting from cold exposures with and without preconditioning. Each treatment consisted of 9 samples of 20 CLB's per sample .................. Mortality levels resulting from various preconditioning treatments. Each treatment consisted of 9 samples of 20 CLB's per sample. Means followed by the same letter are indistinguishable at the .01 level ................ Mortality level resulting from various preconditioning treatments. Each sample consisted of 20 CLB's .......... The effects of preconditioning on mortality levels from various cold exposures ...................... A comparison of preconditioning and recovery in reducing mortality caused by a standard exposure. Means followed by the same letter are indistinguishable at the .05 level Recovery from exposures at -5° (F) ................ Recovery from exposures at O0 (F) ................ Determining the amount of recovery duging a repeated sequence of exposures of 1 hour at -5 and 15 minutes at 0° (F) ............................ Stem densities and relative amount of resistant wheat in foliage samples on different dates .............. Total eggs found on suscepgible and resistant plants in 5 foliage samples of 1 ft. in each plot ............. Adult feeding damage (mm. removed/ft.2) on resistant and susceptible wheat in each plot .................. Adult cereal leaf beetle densities per 1000 sq. ft. as determined by a sweepnet survey ................. Cereal leaf beetle larval densities per 100 square feet as determined by a sweepnet survey ................ Observations on flight distances and intervals between flights for cereal leaf beetles in susceptible and resistant wheat ......................... Cereal leaf beetle egg and larval densities and total emergence trap catch in each of the 4 square-foot sub- plots (densities are per square foot) .............. viii Page 119 121 122 125 127 128 147 148 151 Table C8. Evalu 4-ft. at ion of cereal leaf beetle pupal cells in the subplots ......................... ix Figure A1. A2. A3. A4. A5. A6. A7. LIST OF FIGURES Cumulative emergence of CLB's as a function of degree- days > 48° F for 4 years' data from Gull Lake, Kalamazoo County, Michigan ................... Emergence rates of CLB' s as a function of degree- days > 48°F for 4 years' data from Gull Lake, Kalamazoo County, Michigan ................... Adult CLB's in an 1842 acre study area as measured by a sweepnet survey throughout the springs of 1971-1974. Calculated densities are determined by fitting an ex- ponential decay to the declining phase of the regional densities ...................... Regional densities as measured by a sweepnet survey compared to densities simulated by applying annual mortality rates to emergence cage results ...... Two strip spray patterns showing a strip comprised of a sprayed and an unsprayed portion ........ The relationship between exposure time of CLB's on treated plants and mortality at various times after applying the insecticide to the plants ........ Insecticide exposures required for various mortality levels as the insecticide decays in time ....... Insecticide concentration as a function of time (in hours daylight) ................... The relationship between exposure levels and mortality Mortality resulting from applying 2 equivalent doses of insecticide to CLB's with different time intervals between applications ................. A comparison of a field experiment and a computer simulation involving spraying 8.5 ft. of 20 ft. strips with malathion at a rate of 1.5 lbs/acre ....... Page ...... 22 ...... 23 ...... 67 ...... 70 ...... 85 ...... 91 Figure Page A8. Results of computer simulations showing the relationship between the width sprayed of a 40 ft. strip and the resulting final mortality level for various diffusion 9 rates .............................. 2 A9. Results of computer simulations showing the relation- ship between the width sprayed of a 40 ft. strip and the resulting final mortality level for various initial insecticide concentrations .................... 94 81. Mortality levels resulting from continuous exposures of CLB's to various temperatures ................. 109 82. Hours exposure required for 50% mortality (LT50) at various temperatures ....................... 111 B3. A linearized relationship between LT50 and temperature where the y-axis has a logarithmic gcale and the x-axis is converted to (25° - Temperature) ............... B4. The relationship between exposure level and mortality ...... 115 85. Mortality resulting from an exposure of 3 hrs. at -5° following preconditioning treatments at different times and temperatures ......................... 123 86. Recovery levels as a function of time and tempsrature of recovery from initial exposures at -5 and 0 F ......... 133 87. Rates of recovery from -5 and DOE as a function of recovery temperature and 2 possible alternatives for a general function relating the rate of recovery per degree of temperature increase to the initial tempera- ture exposure .......................... 135 C1. Layout of the resistant wheat plots studied showing the location of the resistant (100% R) and suscepatible (0% R) plots and the plots containing mixtures of R and 5 seed ..... 141 C2. Densities of CLB adults in pure resistant and pure sus- ceptible wheat throughout the season. Densities of eggs in wheat and adults in oats are included as a reference ..... 144 C3. Larval densities in the pure and mixed plots of resistant and susceptible wheats throughout the season ........... 152 C4. Within generation survival of CLB's in pure resistant and pure susceptible wheat as a function of initial egg density . . . 158 xi PREFACE The basic theme of this thesis is concerned with the behavior and survival of populations of cereal leaf beetle (CLB) adults. To this end, I conducted field research at the Gull Lake Research Farm between 1972 and 1974. Basic objectives included determining where beetles overwintered, how they moved around in the spring and summer, and the rate of survival throughout the year. Insofar as I shared these objectives with my predeces- sor on this project, Dr. William G. Ruesink, who used similar techniques in gathering the same types of data, I have incorporated much of his data, taken from his Ph.D. thesis and field notes, and integrated them with my results. This allows a reevaluation of his data in light of subsequent work and provides a more complete data set on which to base current inter- pretations. Rather than crediting every individual table or estimate which was taken from Ruesink's work, it should suffice to say that all data col- lected prior to 1972 was taken by Bill. Other researchers are referenced whenever their work is used. In the course of conducting and analyzing this field work, 3 special areas of interest and research opportunity developed. These projects, described in Appendices A-C, are all outgrowths of the basic study on be- havior and survival. Since in all 3 cases, the techniques, analysis, and discussions are very different from the basic field work, these research efforts were each written as separate sections. The first section discusses field work at Gull Lake concerned with the behavior, survival, and distribution of adult cereal leaf beetles. Early observations on adult movement and susceptibility to insecticides led to the idea that strip spraying might be developed as a control measure against the CLB. This concept is developed in Appendix A, which discusses the field and lab measurements, the theory, the economics, the computer simula- tions, and the field results of strip spraying for CLB control. Field ob- servations made for this project also contributed to an interpretation of the field densities measured at Gull Lake. One of the more important and least understood aspects of CLB dynamics is overwintering mortality. A model predicting overwintering mortality could be an important component of an on-line system for CLB management. For this reason, a fairly comprehensive set of lab experiments was conducted to determine and model the response of CLB's to low temperatures. This work, reported in Appendix 8, though not complete in all respects, allows a new insight and some new dimensions to the subject of overwintering survival, a topic which has not received anywhere near the attention it deserves in entomological research. The third appendix discusses an evaluation of the first large scale planting of wheat resistant to the cereal leaf beetle. The orientation in this project is toward determining the impact of the wheat on the beetle and its parasites, and the probable outcome of releasing resistant wheat in a control program. As with each of the other projects, this effort pro- vided additional insight into the behavior of the beetles; but, probably to a greater extent than with the other chapters, the interpretation of these results was influenced by the results of the other Gull Lake field work. INTRODUCTION The life cycle of the cereal leaf beetle (CLB) in North America has been reported by Castro gt_gl, (1965). The within-generation population dynamics was reported by Helgesen and Haynes (1972), and the interaction with parasites was described by Gage (1974). In modeling the dynamics of CLB populations, certain new information has been required on the between-generation dynamics. Initial field work allowing preliminary estimates for these parameters was conducted during 1970 and 71, and was reported by Ruesink (1972). This thesis discusses subsequent field and lab work intended to refine estimates of these param- eters and includes the results reported by Ruesink (1972) as well as addi- tional results of 1972-74. The cereal leaf beetle, a recently introduced pest of small grains, overwinters in the adult stage in Michigan. In the spring, adults feed and oviposit on a variety of grasses, but are found in greatest densities in small grains where subsequent larval feeding can seriously damage the crop. The adults have been found overwintering in a number of protected sites, generally at or near ground level. Castro (1964) found overwinter- ing CLB's under the bark of trees, in logs, in folded leaves, in straws on the ground, in corn stubble, in bailed hay, farm structures, beehives, field margins, and in woodlots. In areas of high CLB densities, adults are fre- quently found overwintering in large numbers in grain stubble, especially where the stubble field borders a woodlot or dense fence row to the north and/or east. Overwintering mortality was measured near Galien in southwestern Michigan by Castro (1964) at 100% in caged beetles exposed 4 and 12 feet above ground. Beetles held in cages at the ground surface experienced 68% mortality in the winter of 1962-63, and 48% the following year. Denton (1973) measured adult CLB mortality in the same area during the winter of 1971-72 and found that by March 28, 1972, adult mortality in standing wheat stubble averaged 69.9% vs. 49.1% in prostrate stubble. Yun (1967) and Wellso g__al, (1970) recorded the mortality rate of beetles stored at 38°F in the lab during the winter. Both reports noted about 4% mortality during the first 2 months in storage and a gradual increase in mortality rate after that time, such that by the end of March there was about 82% mortality observed by Yun, and about 75% observed for two consecutive years by Wellso gt_gl, Castro (1964) noted apparent predation on beetles in moist overwintering sites as evidenced by a high incidence of body fragments and low density of living beetles in these sites which had high numbers of predaceous insects. Following emergence from overwintering sites, the movement of spring adults has been assumed by numerous authors to have a rather rigid chrono- logical order: spring grasses, to winter grains, to spring grains. In an experiment which involved spraying all the wheat fields in a township to kill adult CLB's and measuring the impact of this treatment on CLB den- sities in oat fields, Wells (1967) noted no decline in populations in oats as compared to controls. Despite some complicating factors, his results cast some doubt on the accepted sequence of movement. Gage (1974) found densities of egg and larval stages of cereal leaf beetles to be related to planting dates of wheat and oats. Late planted wheat and early planted oats had higher densities than the normal plantings of these crops, indicating further complexity in the behavior of the adult beetles. METHODS The Study Area An 1842 acre area in the northeast corner of Kalamazoo County, Michigan, was chosen for this study; the majority of that acreage belongs to Michigan State University Kellogg Biological Station. For the purposes of estimating the total number of cereal leaf beetles in this region, the 1842 acres were divided into several categories, then density estimates were taken from each category. The 1327 acres under cultivation was distributed among about 300 fields ranging in size from .8 acres to 35.8 acres. The remain- ing 515 acres was subdivided as follows: woods, 262 acres; fence rows, 13 acres; roadsides, 27 acres; and others, 213 acres. The final category con— tains such things as lakes, roads, buildings, and lawns. None of these were sampled as they were considered unsuitable as habitats for the cereal leaf beetle. Overwintering Sites An extensive search was conducted over a 7-year period to find the preferred overwintering sites of the cereal leaf beetle. Three basic methods were used to determine overwintering sites. During the summer and fall of 1969, samples of 3 square feet were dug to a depth of 3 inches and all plants and soil in these samples were run through a cotton gin trash mill. This machine was acquired from the Plant Pest Control division of the U.S. Department of Agriculture where it had been designed and used to survey for pink bollworm larvae in the trash left from ginning cotton. The machine consists of 2 revolving screen cylinders which sift the ma- terial of the sample into 3 parts according to particle size (Curl and I White, 1952). When the soil was loose and dry, this machine efficiently separated the beetles from the soil and most of the debris. Excessive moisture caused mud to clog the screens, so the beetles were not then separated out. For this reason, this technique was not used in the spring. Especially designed emergence traps (Gage and Haynes, 1975) were used from 1971 to 1975 to sample the number of beetles emerging, and the rate of emergence, from overwintering sites. These pyramid traps covered a square yard of ground surface and caught emerging insects in a pan of ethylene glycol when they reached the t0p of the screen sides. Beetles were also found in their overwintering sites by direct obser- vation in 1970 and 71. Old fence posts were torn apart, bark was stripped from wild grape, and leaf litter was sifted in the field. These latter techniques did reveal some beetles, but, in general, the gin mill and emergence cages provided the most information. Cage Studies of Adult Mortality In 1970, the mortality rate of adults was studied using 6.6 ft. square (1 milliacre) x 6 ft. high plastic screened cages. These cages were equipped with a zipper door and open bottoms with plastic flaps to be buried so that beetles could be confined to host plants in the field. Two cages were used for spring adults: during May they were in wheat, and in June they were moved to oats. Four cages were used for summer adults: 2 in oats and 2 in corn for the first 3 weeks of July. In every case when a cage was first set up in a new location, it was necessary to remove the resident beetles before the study began. This was accomplished by a visual search using a hand aspirator to collect every beetle seen. When no more could be found, the person left the cage for about a hour and then repeated the search. Normally the second search caught about 1/10th as many beetles as the first. Each week 250 beetles were put into each empty cage. After 6 to 8 days the cages were again emptied using the same search process described above. When the beetles were introduced into an emptied cage at the start of each trial, the jar containing them was opened and placed inside the cage. Those found dead in the jar when the cage was emptied a week later were subtracted from the number introduced before computing mortality. In 1971-74 a different type of cage was used to evaluate adult CLB mortality in the field. Since beetle densities in the study area were greatly reduced, it was decided that the large cages were too inefficient. Smaller cages, with fewer beetles were used, allowing more replicates, and statistical comparisons of survival rates in different crops. The cages constructed for this purpose were made from plastic screen formed into a cylinder 12 inches high by 3 inches in diameter. These were at- tached to a small stake and fitted with l-inch thick foam rubber end- pieces which were slit to allow a plant to be inserted through a cage. Ten beetles were placed in each of these cages for a few days, and at the end of the exposure period the dead and alive beetles were counted. These small cages had the advantage of minimizing microclimatic effects and allowing a determination of whether any beetles had escaped. Population Survey In 1971 and 1972, each grain field in the study area was sampled to determine the number of adult beetles in that field. Each of the approxi- mately 50 fields were sampled at regular intervals to determine changes in the populations of both spring and summer adults. In 1973 and 74, about 50 non-grain survey sites were added to determine the distribution of beetles throughout the study area. Mortality rates were determined from the survey data by determining the rate of decrease of the regional population each year. The sampling techniques used in the survey varied with crop height and beetle density. In 1971, grain less than 10 inches tall was sampled using a thrown stick technique, while taller grain was swept with a 15- inch diameter sweepnet. One sample with the stick technique consisted of: 1) throwing a 12-inch garden stake at least 10 feet; 2) moving the stake 2 stake lengths further down the grain row; and 3) counting the beetles in 12 inches of 2 adjacent grain rows. One sample with the sweepnet tech- nique consisted of 10 sweeps, each 5 feet long, keeping the top rim of the net as close as possible to the top of the grain plant. Sweepnet catch per sweep (C) was converted to number per square foot (0) by the equation given in Ruesink and Haynes (1972): D = c (0.20 + 10K) where K = —.06 + .02 H - .017 (T + 105) + .661 log]0(W+l), H = grain height (inches), T = temperature (OF), S = solar radiation (cal/cmZ/min), and W = wind (mph). 10 Most of the needed weather data were available from a weather station set up within the study area; however, some data came from U.S. Weather Bureau records for Jackson, Michigan. In 1972, because of reduced CLB densities, the number of sweeps in a sample was increased from 10 to 20 and, in short plants, rather than count- ing the beetles in 2 linear feet, a 30 linear foot sample was used. In 1973, a 25 sweep sample was used, and in 1974 the sample was increased to 50 sweeps. In both years, the 30 linear foot sample was used. In all 4 years, 4 samples were taken in each field on each sample date. RESULTS Direct Observation of Overwintering Sites On November 6, 1970, an old weathered fence post was torn apart: 18 live and no dead beetles were found in its cracks and crevices. This sample indicated that significant numbers of beetles may overwinter in micro-habitats which are difficult to quantify. In early April of 1971, additional fence posts, logs, etc., were examined to determine if large numbers of beetles successfully overwintered in such habitats. 0f the 162 beetles found in 4 old fence posts, a decay- ing stump, and under wild grape bark, only 9 were alive. Since these samples were taken before the weather was warm enough for spring emergence to begin, the difference between the observed survival in November and in April represents overwintering mortality. In November 100% of the beetles were alive, while in April only 6% were alive; therefore, overwintering mor- tality in above ground exposed habitats for 1970 is estimated at 94%. Gin Mill Samples The gin mill samples taken during August 11-18, 1969 (Table 1), showed CLB's distributed among all habitats sampled. The beetles were found to be distributed throughout croplands at a mean density (weighted by sample size) of 5,740.4/acre compared to 18,754.9/acre in non-croplands. An analysis of individual samples indicated that beetles were randomly distributed in croplands. A chi-square test indicated no significant departure (P<.75) of the observed densities from a poisson distribution. 11 12 m 0 ON 0 Now moo.m m m mm mama: e o mp C Pan mam.m~ Fm Am AN mmu?muaom o 0 mm e mow ome.Fm m m, mp agom mucwa m o m _ mPN.~ ome.m NF A New muooz mucmFaoLuucoz m, N we _ NmN.P oem.¢ m N New ccou NP 0 mm 0 men mpo.F m P QFN acFac_< m o om o ooo.~ mom.“ Fm F_ mom apnnzpm grace a 0 mp F PmN.m NNN.@ A m mum m_nH macmpmocu mammamm m.muu mm_q5am m.muu Am.ooo._v acu<\m.msu ma_asmm ”.muu macu< Smyrna: ”.moo Payee o .>oz-m~ .puo .puo-.pamm m_-__ pmzmz< .mmm_ to __m$ ucm guessm mum, c? P_rs smog“ arm any cw ummmmuogn mmpaEmm new» magnum-» cw ucaow mupanm mpummn mmm_ Fumemo ._ a_nac 13 Furthermore, the coefficient of dispersion (C.D. = variance/mean) of 1.149 is close to the 1.0 expected for a random distribution. In the non-croplands however, beetles are not randomly distributed. A 0.0. of 4.568 and a highly significant chi-square indicate a significant clumping or aggregation of beetles compared to a poisson distribution. The subsequent samples taken from late September to early November indicate a large decrease in beetle density from that determined by the August samples. 0f the 161 samples taken in croplands during these 2 sample periods, only 4 CLB's were found (.075/yd2) compared to the 1.186/yd2 ob— served in August. A similar decrease was observed in the non-croplands where 5 beetles were found in 77 samples (.l95/yd2) compared to 3.875/yd2 in August. Clearly, the beetles had moved between the late summer and fall samples. This movement is best interpreted in light of overwintering results and, hence, this aspect is deferred to the discussion. Emergence Traps From 1971 to 1973 l yd2 emergence traps were set up each spring in the Gull Lake study area to determine where beetles overwintered and the rate of spring emergence. These traps were not placed at random in the environ- ment, but were placed in categorized habitats listed in Table 2. For the purpose of determining overwintering emergence, the woods in the study area was classified as either dense (mature trees with a litter ground cover), or sparse (open with grass ground cover). The perimeter of each woods, including 20 feet into the woods, was classified as woods edge. The results of these 386 samples are included in Table 2. Since CLB densities decreased in each of the 3 years, it is not possible to make direct comparisons be- tween yearly catches. Furthermore, since there are clearly differences in 14 uo—asmm go: 1 - "mpoz ooo.ooH mmo.m mHN.N¢ mNH.mm eNm.HH Nee. .M umpmznn< mNm.¢m NmH.m Noo.o¢ mmm.mm Nom.~H mac. .M uwucmwmz mHm.mH HHN.¢m mm¢.¢m mmo.m~ 11 - - - -1 Punch mo & omm.H NHN. mmq. mmc. QON. - 11 1- - 1- Nu>\xpwmcmo mN mm mm 0N - -1 - - 11 mange m on NN a 1- - - 1- - m.m40 mNmH mom.N omm.mm Hum.mN cmN.m Punch N0 N mmN.m moH. CNN.N moo.H mam. ooc. coo. oco. ooo. coo. Nu>\apwmcmo mm cm 3‘ a s 2 m m: S 89; e we Ne oH o o o o o m.m40 NNmH Nmm.mH mm~.m moH.oN eNH.¢ Noo.N Pogo» we a omN.¢N oom.¢ omN.H ooo.N~ coo.H oom. mmm. -1 ooo.N mNm. Nu>\>prmcmo w v e N NH m - H m mace» mm m mm N o H - N m m.m4u HNmH Ponce mzom wmcm mvooz mcooz mucmmmocu opossum mumNP< cgou mPuH mucmu muooz mmgmam mmcma Lee Pouch mucmpaoLu-coz mccmFaoLu .mgmgp mucmmcmsm use» mgmzom-H mcvm: mNmH new HNmH comzumn mxmg Ppsw pm mmpwm mcwgmpcvzgm>o soc» ocwmcmsm unused mupzum oFummn mmmF mecmo .N mpnme 15 overwintering densities among the habitats, and unequal efforts went into sampling habitat types each year, it is not very meaningful to directly compare total CLB catch per year. Comparisons between years were made by determining yearly density/yd2 of CLB's in each habitat type, summing these densities, and determining the % of this total that came from each habitat type. This % of total distribution presented in Table 2 is inde- pendent of sampling effort and regional density and, thus, the 3 years' results are comparable. The average distribution for the 3 years was de- termined by weighting the yearly distributions by the number of samples per habitat, and this distribution was then adjusted to total 100%. On the basis of this adjusted mean % distribution in Table 2, it appears that beetles successfully overwinter in the highest densities at the edge of woods (42.8%). Sparse woods is a slightly less favorable habitat (35.2%), followed by dense woods (11.9%), fence rows (9.6%), and croplands (.4%). To determine the generality of this distribution in an area with a different CLB density and different topography, a total of 90 emergence cages were placed near Galien, Michigan, in 1974 and 75. These results (Table 3) were analyzed in the same manner as the Gull Lake samples, giving a reasonably similar distribution of emerging CLB's. The primary differ- ence was a much higher density in fence rows (36.1% vs. 9.6% at Gull Lake). This difference in density reflects a difference in the types of fence rows-~those at Gull Lake consisted of mostly shrubs with an occasional large tree, compared to more mature and dense trees in the sampled fence rows at Galien. Since, except for fence rows, the distribution of CLB's at Galien was quite similar to the Gull Lake distribution, a mean distri- bution was calculated for the 2 sites by weighting the estimates by sample 16 umHasmm Ho: 1- "muoz coc.ooH NHH.om N¢¢.©N omo.mN mNm.m mum.m .M nmpmzhv< www.moH oem.nm cme.NN mNo.mN mmm.m NHm.m m.umucaHm3 omN.mH oN¢.¢N Hmm.mm Hm¢.Nm mou.m Hmpo» to a mmw.¢ mmm.m NH¢.m NHm.H -1 1- Nu>\>HHm:mo NH NH NH NH - - munch mm mm mm mN 1- - m.mHu mNmH HoH.Nm me.mH HNo.mH ch. NHm.m Hmpoh we a Nm¢.m ON¢.m oom.H mN¢.H com. mmm. mmm. Nu>\HHHmcma m m N m m m mamch He m 0H a H H m_mHo cmmH ammo» mzom ovum mcooz muooz mucmHaogu mHnnzpm mucmu moooz mmgmam mmcmo Low Hmuop mucmHoocu-coz mccmHaogu .mamgp mucmmgmeo tea» wgmacm1H mcwmz mNoH ucm ean mcwgzu .cmmrson .cmHHmo Lam: mmuHm mcngch3Lm>o ease mchLmem Hcmzmu mszum mHHmmn emmH Hmwcwu .m mHan 17 size and adjusting the resulting distribution to total 100%. This weighted mean distribution for 5 years' data from 2 sites shows beetles emerging from woods edge in the highest density (40.4% of the total). Sparse woods is slightly less favorable (33.4%), followed by fence rows (15.0%), dense woods (11.1%), and cr0plands (.6%). Distribution in 0verwintering,$ites The results of the 5 years' individual emergence trap catch are sum— marized in Table 4. To determine whether the beetles were randomly distri- buted, a poisson distribution was fit to each of the 5 data sets (Table 5) using the mean number of CLB's per trap to determine the distribution, and the observed and expected distributions were compared with a chi-square test. Each of the 5 years' results turned out to be significantly differ- ent from a poisson distribution. Since the coefficient of dispersion was greater than 1 for each year, it is apparent that the beetles are aggre- gated in their overwintering sites. An inspection of the observed and expected frequencies reveals that in each year there were more zeros and more large numbers than would be expected in random distribution. This is probably a reflection of the differences in density among the different habitat types, environmental heterogeneity within habitat types, and pos- sibly a tendency for beetles to cluster in overwintering sites. Rate of Emergence From Overwintering During the 3 years that natural emergence was measured at Gull Lake, the traps were checked at frequent intervals to measure the rate of emergence from overwintering sites. Additionally, in the fall of 1973, 15 sample sites were selected in sparse woods, dense woods, and idle grass fields (5 sites/habitat) and 2 square fiberglass screened envelopes with 18 Table 4. Densities of cereal leaf beetles caught emerging from overwinter- ing sites at Gull Lake and Galien, Michigan, in each of the 1- square yard emergence traps. FREQUENCY OF OCCURRENCE Gull Lake Galien CLB'sZIrap 1971 1972 1973 1974 1975 O 12 148 129 14 16 1 6 22 23 5 12 2 5 7 10 4 3 2 3 3 4 3 4 2 2 1 2 6 5 2 2 3 6 1 3 7 1 2 2 8 1 l 9 10 l 1 11 1 2 Additional Trap Catches (each number occurred only once) 16 12 15 65 15 16 30 17 20 24 25 31 Total Traps 30 188 168 30 60 Tota1 CLB's 117 127 66 68 288 X'CLB's/Trap 3.90 .68 .39 2.27 4.80 s2 142.54 7.45 .76 10.41 47.25 C.D. (E %3. 36.58 10.69 1.97 4.59 9.84 19 1184.4N 1 Nx Nm.H oH ax Hm o em.H HH 0 m m¢.m H m m~.m N N kth.mH u Nx mm.m m m m¢.oH m m o¢.N m wk mm.m N_fl Nm.oH o a m¢.m N e oH.m m m eo.m e m mm.m e N mm.N o N mdeNmN Nm_HNH H . 3N m H 3. 2 o 3 2:3 2n: o mmogm yams; .mcmo 14cmgm «mag; .mcma .axm .mno .axm .mno mNmH ¢NmH «kmw.mv u Nx om.H N NA mm.m mmH m_Ho N Nm.N H m seem.mH u Nx eeom.HN u Nx Nm.¢ o m . mm. m mx mm.m N e No.2nNNH 21 NA 8 253 :mm m 8.0 N m mN.m oH N mm.HN N N Nm.¢ m N 3.3 8 H 8.3 NN H om.N_HNm.N N T H N¢.mHH mNH o Nm.mm N¢H o Hm. NH 0 .amgm .amgu .mcmo .amgu «mag; .mcmo .Iwumgu «mogm .mcmo .qu .mnc .axm .mao .nxm .mno mNmH NNmH HNmH mmHucmaamgH um>cmmno .coHuaanpmHu commHoa m sogH umpumgxm mmocu o» mcHLmanu mmguumu amga mucmmgmso HmavH>HucH Ho memo» m we mHmHan< .m oHan 20 25 adult CLB's in each were placed in each site at the soil surface. The rate of emergence of these beetles after cutting open the envelopes in the spring of 1974 is included in Table 6. In Fig. l, the results in Table 6 are plotted on a log-probability scale, omitting the 8% of the beetles that emerged after May 18 in 1971. The relationship between cumulative % emergence on a probit scale and the log of degree-days > 48°F is quite linear. Regression equations fit to this transformed data gave an average r2 of .968. The degree-day values corresponding to 50% emergence for 1971-74 are 121, 83, 135, and 123 oD>48, respectively. These occurred on April 21, 28, 17, and 22 in the respective years. From the regression equations fit to the data in Table 6, the rate of emergence was determined for each year. These results, pre— sented in Fig. 2, show that for 3 years the emergence curves were very similar; however, in 1972 emergence peaked earlier and was completed sooner on a degree-day scale than in other years. It is interesting to note that this year was relatively cool and, although the peak was earliest on a degree-day scale, it was the latest of the 4 by calendar date (April 26). Peak emergence at Gull Lake from 1971-74 occurred between 75 and 125 degree- days which corresponded to April 18-26. Regional Emergence at Gull Lake To determine the number of cereal leaf beetles emerging from overwinter- ing in the 1842-acre study area, the mean distribution for the 3 years of Gull Lake results was used. Since in each year the number of traps used and beetles caught was relatively small, it was felt that the average of the 3 years was more representative of a single years' distribution than the density actually measured. This is particularly true in 1972 and 73 when no beetles at all were found in croplands. 21 mvm NNm mH Nam mmN NH Hem oNN vH Nmm HmN NH omm ch m Nmm N¢N o omm mmN e ONm mHN N Nmm HHm NON om mom mNH mN NNN NwH oN mON cmH eN NmH CNH HN mN mo oH NN mm NH Ne mm mH Nm mN mH H cc m H Hm N HHLm< .Em m¢xoo mums .an ¢NmH mo ¢¢m ma mo mNm “H mm mmN m \w mo #NN N \H .51: co NHN NN mm cmH MN ¢¢ nmm 0N\ NN NHH NH 5 Ha“ .EN maxoo oumo .Ezu mnmm NNH mON Mm mNH mmH m mNH hmH m NHH HmH m \e OHH mNu N \H 8H moH mm mN mm on NN mN Nu mN\¢N NN Hm NN\HN NN om ON\mH c we 1mg . Em w¢AQO $900 .530 NnaH NHH moo m mHH mom m oHH Nm¢ H mesa mHH NHe mm NHH 0mm mN on N¢m NN on mmN wH moH ocN mH mm mmH oH mm ¢mH N NN meH m Nam No ¢mH om Nm eNH HN cH mm NH NH mN mm HHL < .Em wexoo mums .an HNmH .mmmmu mucmmgmso figs» mgmaam1H an umgzmmwe mm meHm mcHgmch3Lm>o eoLH muHaum mHHmmn NmmH Hmmgwu No mucousmEm w>HpmH=s=o .o ancH 22 99.9" 99* 95- m 80- 0 Z w 60* 3:” m 40" 2 DJ - 20 1971 °\° 1972 1973 5" 1974 '— 1 l {.1 l 1 1 l 1 a 20 40 60 ICC 200 600 DEGREE DAYS > 48°F Fig. 1. Cumulative emergence of CLB's as a function of degree-days > 48°F for 4 years' data from Gull Lake, Kalamazoo County, Michigan. (Neg. 752244-12) 23 HHN-chNmN .mazv .cmmHgon .Nuczoo ooNNENHax .mme HHao eogH many .mgmw» e No; home A maov-mmgmmn No coHuucam a ma m.m4u No nouns mucmmgmeN .N .oHN momHvA m>V mmHHHmcmo umumquHmu OOH ome.m NHN.NH NNH.mN eNm.HH Nee. HNV coHpanaumHo N ow oH mNH mNH NNm.H mogu< Hmpoh mmnHmvmom mmum mcooz muooz mncmHmocu new mcooz mmemqm mmcmo m3om mucmm mN48 %[OD 1972 May 10-12 Wheat 12 108 5.132 19 .553 24-26 Wheat 11 89 5.660 22 .528 June 5- 7 Oats 4 86 2.247 38 .120 1973 May 7-10 Wheat 32 169 5.616 42 .412 11-15 Wheat 15 175 2.035 16 .513 21-25 Wheat 17 103 3.747 54 .282 June 4- 8 Oats 22 158 3.207 86 .151 11-14 Oats 29 141 6.044 71 .263 21-25 Oats 27 43 11.469 74 .656 1974 May 2- 9 Wheat 11 69 2.091 26 .567 9-15 Wheat 1 149 .111 34 .020 16-22 Wheat 6 154 .635 64 .060 22-30 Wheat 2 148 .168 101 .013 June 1- 7 Oats 12 158 1.213 125 .059 7-14 Oats 45 135 4.030 125 .230 22-28 Oats 30 90 4.680 89 .323 29 Mortality rates in different crops were compared simultaneously by using several cages with 10 CLB's per cage in each crap. These results (Table 10) indicate no difference in survival among the several pairs of crops tested, except between wheat and oats (beetles survived longer in oats). This difference between wheat and oats may be real or, in light of the other results, it might be an artifact of sampling: for example, it is possible that some pebbles got into the net when sweeping the lO-inch oats. In each of the 4 years' caged mortality studies, the cages were moved from wheat to oats the first week in June. A comparison of the mortality rate in wheat to that in oats the following week reveals that in 2 years there was a greater rate of mortality per day in wheat than in oats, and in the other 2 years the reverse was true. On the basis of these results, it appears that the mortality rate of spring adults may be the same On all hosts. Population Survey The weekly sweepnet survey results (Tables 11 and 12, and Fig. 3) show the densities of adult CLB's in the 1842-acre study area in the springs of 1971-74. In the first 2 years, only the croplands were sampled; however, in 1973 and 74 all representative habitats were sampled. In summarizing the sample results, the density in each of the grain fields was multiplied by the size of the individual field and similar crops were then added together. In fields which were not sampled and in all non-crop areas such as fence rows, 1Pastures, idle land, etc., an average density was determined for the habitat type and multiplied by total acres of suitable grasses within that habitat type (only the grass component of these habitats was deemed suitable for CLB'S and for sweepnet sampling). For example, roadsides were estimated to be 70% grasses, fence rows 85%, and idle fields were individually estimated for Percent composition of suitable grasses. 30 Table 10. Comparisons of caged adult mortality rates on different crops in 1973. 7' % Standard Significance Date Crop Mortality Deviation Samples Level May 5-10 Quack grass 13.000 18.288 10 N.S. Wheat 16.000 17.889 20 May 11-15 Rye 12.000 12.293 10 N.S. Wheat 17.000 13.375 10 May 21-25 Quack grass 7.895 11.343 19 N .5. Wheat 14.167 11.645 12 June 4- 8 Oats 12.222 13.086 18 .05 Wheat 24.737 20.377 19 31 Table 11. Seasonal density of adult cereal leaf beetles measured by a 1971 0 Winter Spring Date D>48 Grain Grain --Overwintered Beetles-- May 5 149 1,571,000 N.S. 10 195 1,091,000 N.S. 19-20 321 716,000 6,380,000 22 342 620,000 5.601.000 26-27 387 873,000 5,416,000 29 412 339,000 2,873.000 June 2- 3 466 64,000 814.000 5 535 96,000 1,108,000 9-10 620 66,000 96.000 12 679 17,000 146,000 14 727 80,000 184,000 17-18 801 9,000 61.000 22 935 12,000 26,000 --Summer Beetles-- June 25 1,020 42,000 502,000 28 1,115 101,000 780,000 30 1.185 380,000 1,360,000 July 2 1,232 404,000 940,000 5- 6 1,307 263,000 856,000 8 1,391 0 459,000 10 1,444 2,000 422,000 sweepnet survey in 1971 and 1972. 1972 0 Winter Spring Date D>48 Grain Grain --Overwintered Beetles-- April 19 69 O O 28 88 98.505 0 May 5 144 582.306 0 11 181 449,668 4.208 18 270 124,418 121.067 25 422 142,660 140,983 June 2 544 71,136 134,882 14 773 12,142 39.714 22 909 3,601 23,680 -—Summer Beetles-- July 4 1,110 172,260 476,469 7 1,146 232.997 1.443.086 11 1,243 23,881 1,165,060 13 1.299 35,497 1.877.786 17 1,399 33,204 1,111,166 21 1,517 21.898 161.320 24 1,616 9,468 225,381 - (\MH H «5:0: 0 \A,r -, p... - >ao>11~jnm 1a .~.15.—.-ouz.uc P. \IAN urns-.an-unt.: '1'... which .5 mtohnfi L EMU— h tau-bah s h u -§if1 \ uh \J \ s lufheiuu u. tubaiuiE-EI-u 1,- ~ 1“ -,\-~ 32 em HH -- HH N HH NH H HH NN NNHNENN NNN.HN o -- o o NNN.N o o NNN.NH NNN.NH NNo.H H NHNN NNN.NN NNH.N -- o o o o o NNN.NH NNN.N NNN 4N ONH.NHN NNH.N -- NNN.N o o o o NNN.NNH NNN.N NNN NH NNH.ooH NNN.N -- NNN.oH o o NNN.N o NNN.NN HNN.NH NHH NH NNN.NN NNN.N -- o o NNN.N NNN o NNN.NN HNH.NN NNN N mean NHo.NNN NHN.HN -- oNN.N o NNN.NN NNN.N NNN.N NoH.NN NNN.NNN NNN NN «NH.NNH.H NNH.NN -- NNN.NN NNH.NH NNN.NN NNN.NN NNN.N NNN.NH NHN.NNN NNN ON ooo.Nom.H NNN.NN -- NNN.NN o Hoo.NN NNH.HH o NNN NNN.OHN.H NNN N Na: NNNH ooH HH N N N HH NH N NH NN NNHNENN Noo.NN o o NN¢.NN o o NNN.N o NNN.NN NNH.oH NNH.H NN NNN.NN o o o o o o o NNN.NN NNN.N NNN ON NHN.NN HNH.N o o NNN.N o HNN.H NNN.N NNN.HN NHN.N NNN NH NHN.NNH NNN.H o NNN.NH NNN.HH o NNN.H NHN.NN NNN.HN NNH.NN , NNN N Hom.NNH NNN.NH o NHN.NN NNN.N o NHN.N o NNN.HN NNN.HN NHN H mean NNN.HNN NNN.HN NNN.N NHN.HN NNN.NN NHN.¢N NNN.N NHo.NN NNN.HN NNN.NNN ONN 4N NNN.NN¢ NNH.oH NNN.N oom.oN NNN.NN NNN.N¢H NNN.NH NHN.¢N NNN.H NNo.oNH NNN NH NNN.NNN NNe.NH NNN.N¢ oNo.NN NNN.NH NNN.NN NNN.N¢ NHN.NN o NNN.NHH NON N Na: NNNH HNHOH NNON Nasal NHNNNHN NHNH NNLNHNNN NNHNNH< NHNNNHN NNN NHNN News: Nexoo NHNN tcm 530 c _. ago mwuHmumom .NNNH cam mNmH :H HN>N3N Hmcgmmzm N Na NNNNNNNE NN NNHHNNN NNNH HNNNNU HHNNN Ho NHHmcmc Hmcommmm .NH oHnNH 33 8000 ’ 5000- 3 ....... 1 1. :1 H4000 - 197' '2). m D ozooo- \ 'fi.‘ 9 507. 90%---99%.-' " T a a a 1 41 a L a 1‘1 ..... Mn‘_n n 14— _2_ V 600' \GALGULATEo .- I” \“:“‘\ 3 400- ,1 .- - 1972; E 200 P I” ".\ uGUIuo:I"~"..‘..s-q::o: ‘ ‘ ‘ - UJ ' ’.50 90% 99% .... .- ~'7"°I:‘T1'..°'.;':-'§-‘.':.7.t.7. -------- "’ +1 I 1 4”" l l 1 L a 1 w— 1 l 411 m a 1 LL. 600' 4 . ‘3' 400- - I53'735 _I <-: 200- U 90% a: 20%. 1 . Lu 0 I600 CALCULATED — ENTIRE REGION 1- ' ~ —- ALL GRAINS 3'200' --- WINTER GRAIN o ’ ----- SPRING GRAIN <1 800’ '974 1 OVERWINTERING . EMERGENCE 4oo ’ P507. 93% 99% __ 19 2.4 ‘I 5 -5‘“‘-5‘ 2o 25 I 5 I0 15 20 25 I APR MAY DATE JUN JUL Fig. 3. Adult CLB's in an 1842 acre study area as measured by a sweepnet survey throughout the springs of 1971-1974. Calculated densities are determined by fitting an exponential decay to the declining phase of the regional densities. (Neg. 752244-23) 34 After emergence from overwintering, CLB densities are observed to decline until the last week of June when the summer generation begins to emerge. The increased densities observed in July result from the new summer beetles which feed for a few weeks before leaving the grain fields for estivation. Adult Mortality From Survey Data The results of the 4 years' population surveys for spring adults can be used to compute mortality rates if 2 assumptions are made. First, it is necessary to assume that for 1971 and 72 the rate of decline of popula- tions in all grains combined is the same as the rate of decline of the regional population. An inspection of the 1973 and 74 results in Fig. 3 shows this to be a good assumption. The second assumption is that the migration of beetles into and out of the 1842-acre region balances out and, hence, any decline in density is due to mortality alone. This assump- tion also seems reasonable considering the size of the region, the diversity of habitats, the densities of beetles, and the relative isolation from other grain-growing regions. In 1972 and 74, adult densities peaked in the fields before emergence from overwintering sites was complete (Fig. 3). Since during this period the field density reflects recruitment from overwintering as well as mortality, the field densities that were measured before 99% emergence from overwintering were not included in mor- tality estimates. The rate of decline of CLB numbers after the peak den- sities following 99% emergence is suggestive of an exponential decay of the type g%-= AN where A is a constant decay rate. By performing a linear regression of the natural log of density vs. time (1n y = XX + b) the decay constant was determined for the years 1972-74, and the % mortality per day 35 was calculated as Md = 1 - e4. The curves fit to the data of 1972-74 in Fig. 3 are the product of the regression equations which gave mortality rates of 4.45, 6.83, and 5.93% per day for 1972 to 74, respectively. Because 11 of the 22 oat fields were sprayed with insecticide in early June of 1971, it was not possible to use the same technique to evaluate mortality for that year. To evaluate mortality in 1971, assumptions were made that all beetles in a field at the time of spraying were killed by the insecticide, and that between-field adult migration was small so that only the beetles in the field at the time of spraying were affected by the insecticide. The 9 fields sprayed between May 29 and June 2 contained an estimated 1,156,000 beetles on May 29; the 2 fields sprayed between June 5 and June 9 contained an estimated 234,000 beetles on June 5. Survival (S) from time t to time t+l was computed from _ Nt+1 + (%)K S - NI: _ (19K X10034 where Ni is the number of adults present at time i, and K is the number killed by pesticide between time t and time t+l. This equation adjusts sur- vival close to what it would have been if no pesticide had been used in the region. Table 13 presents the results of applying the above equation to the 1971 data after averaging the results for each week. The average mortality rates for 1971 in Table 13 are 11.23% per day and .71% per degree-day. In all 4 years, the mortality rates as measured by the survey are higher than the average rates measured in the caged studies. As seen in Fig. 3, the survey mortality rate is based primarily on the period from late May through mid-June when cage studies indicate the lowest rate. This fact indicates 36 Table 13. Spring adult mortality as computed from the 1971 regional population survey. Date Mortality (%) From To Days dd43 Total Per day Per dd43 May 20 May 27 7 61 28.7 4.7 .55 May 27 June 3 7 93 61.1 12.6 1.01 June 3 June 11 8 163 66.1 12.7 .66 June 11 June 19 8 208 72.4 14.9 .62 37 that the difference between the field and cage results is probably under- estimated since the comparison is between the lowest field rate and the average cage rate. To determine the rate of mortality on a degree-day scale, the 1972-74 survey results in Tables 11 and 12 were plotted on a degree-day scale (Fig. 4) and analyzed in the same manner as the calendar date results in Fig. 3. By fitting regression equations to the transformed data, mortality rates for 1972—74 were determined to be .710, .368, and .446 %/degree-day > 48, respectively. As with daily mortality rates, these field rates based on degree-days are generally greater than the results of the caged studies. Survey Results vs. Overwinterinngmergence In the 3 years (1971-73) that both survey results and densities of overwintering adults are available, it seems worthwhile to compare the 2 estimates of regional densities. For this analysis, individual CLB adults are assumed to die at the rates measured by the surveys from the time they emerge from overwintering. Thus, the total emergence estimates (Table 7) are not directly comparable to the survey densities of Tables 11 and 12. To make this comparison, the equations for each year's emergence were solved on the computer in degree-day increments and the field mortality rates ap— propriate to each year were applied to the emerging densities in a simulation. The results of this simulation are shown in Fig. 4. The logistic curves show cumulative emergence of beetles from overwintering. The positively skewed curves show the number of alive beetles in the region throughout the season as determined by the emergence cage densities and the survey mor- tality rates. These simulated regional densities were determined for each survey date and compared to the regional densities from the survey. Since 38 9000' 7000' S) 5000- IN 1000 3000' 3; ES (3 <3 c: c: ”--------- , '972 /’ --- OUML. EMERGENGE I I000 - — REGIONAL DENSITY (SIMULATED) 0' SURVEY RESULTS 600' 200- 600- ADULT CEREAL LEAF BEETLES ( 400- ,"-""°“ 200‘ O O l l l l L l 1 1 £ 200 400 600 800 1000 DEGREE DAYS > 48’F Fig. 4. Regional densities as measured by a sweepnet survey compared to densities simulated by applying annual mortality rates to emer- gence cage results. (Neg. 752244-22) 39 the results of the 1973 and 74 surveys indicated that an average of about 75% of the regional CLB density was found in the grain fields, the densities measured in grains in 1971 and 72 were divided by .75 to determine the regional density on each sample date. The average ratios of simulated regional density to survey estimates for 1971 to 1973 are .226, 1.46, and .328 respectively. These ratios are based on several estimates, all of which are subject to considerable error. Since the survey densities and mortality rates are based on a large number of samples, it is probable that the largest errors are associated with the estimates of spring emergence which were based on small samples and relatively few beetles. Cereal Leaf Beetle Behavior In the 4 years of survey results in Fig. 3, it appears that there is a shift between years of beetles from spring grains (primarily oats) to winter grains (primarily wheat). In 1971 most of the beetles were in spring grains; however, by 1973 and 74 only a small part of the population was found in spring grains. Since the acreages of these crops changed a great deal between years, the 4 years' data were standardized somewhat by determin- 'ing the density per square foot in spring and winter grains at peak densities in each crop for each year (Table 14). The ratios of densities in spring/ winter grains for 1971-74 are 9.13 : 2.00 : 0.73 : 0.84. To determine whether this shift between spring and winter grains is related to the condition of overwintering wheat plants, the average height of wheat for all fields was determined for the date of 99% emergence from overwintering in each year. These average heights for 1971-74 are 19, 13, 12.5, and 20.5 inches, respectively. These heights do not correspond in 4O Table 14. A comparison of peak regional densities of cereal leaf beetle adults in spring and winter grains at Gull Lake from 1971 to 1974. Peak Spring Density Year Cropy Density Acres CLBFS/Acre Winter Density 1971 Spring grains 6,380,000 95.1 67,087 9.13 Winter grains 1,571,000 213.7 7,351 1972 Spring grains 140,983 25.6 5,507 2.00 Winter grains 582,306 211.0 2,760 1973 Spring grains 71,928 47.9 1,523 .73 Winter grains 376,988 180.5 2,089 1974 Spring grains 195,859 38.8 5,048 .84 Winter grains 1,340,309 224.3 5,976 41 any consistent manner with the relative CLB densities in spring and winter grains. Also, there is no consistent relationship between the relative densities in the spring and winter crops and the amount of oats available in early spring. In 1971-74 the percentage of oat fields out of the ground on the date of 99% emergence was 37, 30, 50 and 91%, respectively. Thus, neither the maturity of oats nor wheat seem adequate to explain the apparent shift of beetles from oats to wheat which was associated with a decline in regional density. Effect of Crop Age on CLB Density In 1972 and 73 the planting dates of wheat and oat fields were adjusted to get a range of plant maturities to determine the effect on CLB's. In analyzing the survey results for these years, a regression of CLB densities vs. plant height was performed for each survey date. In 1972 there was initially no relationship between plant height and density in wheat, but by May 18 there was a significant (P > .98) negative relationship between CLB density and increasing plant height. Subsequent samples also revealed lower densities in higher wheat, but the slopes were not significant (P > .90 and P > .70 on May 25 and June 2). In 1973, there was not as great a range in plant heights and there was never any significant relationship between plant height and densities in wheat. In oats in 1972 results were somewhat similar to wheat in that the first sample after all oats were germinated and out of the ground (May 25) indicated an insignificant relationship between crop height and beetle den- sity. Subsequent samples gave a significant (P > .99) positive regression of beetle density on crop height, indicating higher densities in the earlier planted, more mature oats. In 1973 and 74, there were also significant posi- tive slopes relating crop height to beetle density. DISCUSSION Overwintering Newly emerged summer adult beetles, after feeding for a few weeks in early July in grain fields and other suitable grasses, distribute themselves throughout the environment. During the late summer and fall, these beetles gradually move into overwintering sites where they are found in different densities in different habitat types and show a distinct tendency to aggre- gate. The gin mill samples taken in mid-August of 1970 (Table 1) indicated that at that time beetle densities outside the croplands were about 3 times as high as densities in crops. Those beetles in croplands were randomly distributed, but those in non-crops were highly aggregated. Samples taken in late September and November indicated the beetles had left croplands but were also difficult to find in non-crops. The most reasonable explanation for this observation is the aggregation in non—crops which was observed in the August sample, as well as in the 5 years of emergence trap catches. Apparently the beetles in the non-crops continue to move after mid-August so that by November they are more aggregated than in August. In 1970 few of the fall gin mill samples were taken in favorable overwintering sites, and those that were taken did not contain an aggregation. An alternative eXplanation would be that beetles overwintered out of the universe sampled, which could mean out of the 1842-acre area, above ground, or deeper than 3 inches. With diversity of habitats in the study area there is no apparent reason for the beetles to leave the area. The 42 43 high mortality of beetles above ground seems to eliminate the second alternative. The third alternative has been tested by repeated efforts to bury adult CLB's at various depths in a variety of habitat types and measure their emergence in the spring. These efforts have not shown beetles able to emerge in any numbers when buried more than an inch in the soil. Thus, it appears that the beetles did in fact overwinter at the soil surface. The emergence cages used in the springs of 1971-73 measured densities of beetles which successfully overwintered in various habitat types. 0n the basis of 3 years' results at Gull Lake, and 2 years' of similar results from Galien, it was concluded that CLB's successfully overwinter in the highest densities at the edge of woodlots. In agricultural areas such as those sampled, if a total of 100 beetles emerged from equal acreages of habitat types, about 40 would come from woods edges, 33 from sparse woods, 15 from fence rows, 11 from dense woods, and less than 1 from croplands. The number of beetles overwintering in the different habitats in a given region depends on the acreages of the habitat types as well as the density of beetles. The large numbers of beetles which have been observed along the edges of fields of grain stubble would probably have been along the same edges of woods or fence rows even without the presence of stubble, although they would be more difficult to find. One criterion which seems to determine overwintering distribution to some extent is the presence of small crevices into which beetles can crawl for the winter. The abundance of such microhabitats in grain stubble, in broken weeds, in sumac patches, etc., apparently serves to concentrate beetles in specific locations within a habitat type and might be the cause of, or at least a factor leading to, the consistently measured tendency to aggregate in overwintering sites. 44 Because of the non-random distribution of beetles within habitat types, it takes a large number of samples and a large number of beetles to accurately estimate the density in a habitat. This factor is attributed as the cause of the discrepancy between estimates of emerging overwintering populations and populations measured in surveys later in the spring. Thus, the fact that the emergence cages considerably underestimated the spring population twice and overestimated it once, is seen not as a problem with the technique, but as a lack of sufficient samples of CLB densities to allow accurate estimates. Density, Behavior and Mortality, Gage (1974), in summarizing 7 years of pOpulation measurements at Gull Lake, reported that after the spread of the beetle into the area, densities of eggs and larvae increased each year from 1967 to 1969, and decreased each year from 1969 to 1973. Although this population decline has corresponded with the increasing densities of the larval parasite Tetrastichus julis (Walker), similar declines have been observed in areas without parasites and, hence, the decline is not attributed entirely to parasitism (Gage, 1974). Since to date, no one has adequately explained the population decline at Gull Lake, it seems reasonable to examine adult behavior and mortality for a possible mechanism. Only a few of the many sources of adult mortality were measured during 1971-74. The average mortality rate of 9.2% per day as measured in 1971 for the newly emerged summer beetles is a very important component that results in a very rapid decline in adult densities. If beetles die at this rate for the 2-week feeding out period reported by Castro (1964), densities of adults 45 would be reduced by 75% before entering estivation. It is probable, however, that this mortality rate is overestimated since beetles were unnaturally confined to crops at a time when they would normally be moving throughout the environment and congregating in areas where food and micro- climates were most suitable. Yun (1967) and Wellso gt_a1, (1970) report only 3-4% mortality during the first 60 days of lab storage of beetles following estivation. This low rate of mortality is not unrealistic for this period because the beetles clearly cannot continue to die at the rate measured prior to estivation. Overwintering mortality is another important component of total adult mortality. Cereal leaf beetles apparently cannot survive Michigan winters in above-ground habitats. Castro (1964) measured 100% mortality at 4 and 12-foot heights in caged studies. In the stumps, grapevines, and other above-ground habitats examined at Gull Lake in the spring of 1971, mortality was estimated at 94%. These habitats no doubt afforded the beetles some degree of protection from low air temperatures as did the standing grain stubble in which Denton (1973) measured about 70% mortality in 1972. Cereal leaf beetles can tolerate quite well, however, the temperatures they normally encounter on the ground as evidenced by the mortality esti- mates of Castro (1964) of 68% and 48% mortality in 1963 and 1964, and Denton's estimate of 49% in 1972. Of the 750 beetles held in screen packets during the winter of 1973-74 at Gull Lake, 349 were caught in emergence cages in spring, indicating about 52% winter mortality. Thus, it appears that beetles survive in the field at least as well as in lab storage at 38° where mortality averages about 77% by April 1 (Yun, 1964; Wellso gt_gl,. 1970). Hence, cold exposure may not normally be an important factor in 46 winter mortality of beetles at the soil surface. What may be more important is the apparent predation in some habitats noted by Castro (1964). This predation may be a factor involved in the often observed tendency of beetles to overwinter in tight crevices which probably serve to limit predation. There is some evidence that habitats which are too warm may result in physiological depletion of fat reserves as discussed by Denton (1974), and the length of the winter may also be a factor in determining winter mortality. The mortality of spring adults following emergence from overwintering sites is another mortality component which is documented for 1970-74. Since the mortality rates as measured by the cage studies were consistently lower than the rates calculated from the survey, it appears that the cages underestimate the regional mortality rate. This might be due to screening out predators. If the spring mortality rates, as determined by the survey, are compared to the regional densities of beetles in either the same year, or in the next year, no apparent relationship develops. These mortality rates of 11.23, 4.45, 6.83, and 5.93%/day in 1971-74, respectively, seem to neither cause, nor result from the peak measured densities in grains of 6,096,000, 582,306, 408,733, and 1,341,109 for 1971-74. One factor which seems to be correlated with the regional p0pulation decline is the relative densities of beetles in spring and winter grains (primarily oats and wheat, respectively). In 1971 when there was a high regional density, most of the beetles went into oats (Table 14). In sub- sequent years the pr0portion of beetles in oats declined as did the re- gional density. This phenomenon is probably best understood in light of CLB behavior. 47 As CLB's move from overwintering sites, they initially distribute themselves fairly evenly throughout the environment and are found wherever there are suitable grasses for feeding, i.e., alfalfa fields, pastures, idle fields, roadsides, etc., as well as winter grain fields. In wheat fields there is initially no relationship between plant height and beetle density in a field. Likewise there was initially no difference between densities in resistant and susceptible wheat during this time as discussed in Appendix C. During the early spring, beetles apparently continue to move both between and within fields, and by the time oviposition begins (late May), beetle densities are significantly correlated with plant height in wheat fields, and there are significantly more beetles in susceptible than in resistant wheat (Appendix C). At that time, beetles are found in the highest densities in the shortest fields. This was particularly ap- parent in 1973 when 4 different ages of wheat were grown in adjacent strips in one field. Densities of adults were consistently negatively correlated with crop height as were densities of immature stages (Gage, 1974). In viewing the survey results of Fig. 3 and Tables 11 and 12, it is not exactly apparent where the beetles come from that end up in oats. The early season samples, particularly in non-crops, are certainly pushing the limits of the sweepnet model of Ruesink and Haynes (1972) used in determin- ing absolute density, and not a great deal of confidence is placed in these estimates. Some facts are apparent, however, as in 1971 when 7 times as many beetles were found in spring grains as were ever measured in winter grains, and in the experiment of Wells (1967) which showed that spraying all the wheat in a township for adults did not affect subsequent oat densities. Clearly, beetles do not follow a rigid sequence of movement from overwinter- 48 ing sites to wild grasses to winter grains, and then to spring grains. Instead there appears to be a period of mobility during which beetles are generally found in the most preferred host plants. Those beetles found in a particular field on a given day may just be passing through as part of a continuous flow between several habitats. Differential flow rates between crops and within crops at different times of the season can account for the different adult densities observed in the habitats. Those oat fields which have germinated and are out of the ground during this period of mobility get high densities of adult CLB's. However, the late planted oats which are unavailable during this time of beetle movement do not get high densities of beetles. Thus, later in the season there is a positive correlation between height of oats and densities within oat fields. The low densities in late planted oats result from several factors. First, because of the mortality of adults, there are fewer avail- able to move into the late oats. Secondly, since oats is the preferred host for CLB's, it acts as a sink and beetles leave oat fields at a much slower rate than they arrive. Thus, there is relatively little movement of beetles between even adjacent oat fields (this is apparent in the strip sprayed field of Appendix A where the density in the sprayed field decreased dramatically, but the adjacent oat field was unaffected by the spray). The third factor involved is the reduced rate of movement as the season progresses. As seen in Appendix A. the beetles move at a greatly reduced rate after the oviposition period begins. This combination of beetle mobility and preferences serves to explain the densities in crops in a particular year; however, it does not explain 49 the observed differences between years. It was already noted that the shift of beetles from oats to wheat during the 4 years surveyed was not apparently related to the planting dates or synchrony of the 2 crops. This shift might be related to the decline in regional density, however, because CLB's are less productive in wheat than in oats. The fecundity and within-generation survival of CLB's are higher in oats than wheat (Helgesen, 1969; Wellso, pers. comm.). Thus, when a large proportion of the regional population is in oats, that population can increase more rapidly than it could if most beetles were in wheat. In turn, if the proportion of beetles going into oats were determined by regional density, the behavior of the adults could have a profound effect on the regional density. There is some evidence that beetle adults resist crowding. Helgesen (1969) reported a maximum density of 5-7 adults/ft.2 which was never exceeded regardless of the regional density, and he speculated that the audible sound which beetles produce might be a factor in regulating adult densities in a field. It is possible that the crowding of CLB's in high density years results in more beetles leaving wheat and other habitats and ending up in oats. In years of low density this crowding would be less important and more beetles could remain in other habitats. This behavioral feature coupled with the higher produc- tivity in oats could serve to greatly amplify, over several years, what might otherwise be a minor change in population density due to weather conditions, plant synchrony, or some other factor affecting the regional density. It is possible that this shift from oats to wheat could have been one of the factors contributing to the decline of the CLB population at Gull Lake between 1971 and 73. Other factors which apparently con- 50 tributed are the reduced survival of eggs and small larvae in oats, the adverse weather conditions in the summer of 1971, and the incidence of parasitism--all recorded by Gage (1974). Implications for Cereal Leaf Beetle Management In light of the new information on the survival and behavior of adult CLB's described in this section and in Appendices A-C, it is important to re-evaluate CLB management practices. Gage and Haynes (1975) suggest directing control programs against the adults instead of the current practice of spraying for CLB larvae which is detrimental to parasites. The effective use of an adult control program requires a method of predicting or monitoring adult densities and an understanding of the significance of these densities with respect to the dynamics of the insect and its relationship with its hosts and parasites. Secondly, should this density be determined sub-optimal, the economic and ecological impact of a variety of control options must be evaluated. These factors are best evaluated by ecosystem models, of which an assortment is available on many aspects of the CLB ecosystem. Before adult control is effectively incorporated into a CLB management program, new ecosystem models will be needed to evaluate the impact of several factors affecting adult beetles. The first of these factors to be considered in a control strategy is the type of crop planted. Based on the shift of beetles between oats and wheat according to the regional population density, it seems reasonable that to the extent it is economically feasible, the planting of oats should be discouraged during times of epidemic population levels. The acreage normally planted to oats could be planted to wheat instead, there- 51 by allowing a buildup of parasites on the wheat (which will probably go unsprayed) and suppressing a further population increase in the oats, on which beetles are more productive. Growers would save on insecticide costs and maintain parasite populations by precluding the necessity of insecti- cide applications on oats, the "preferred" host of epidemic populations. Oat acreage could be increased during years of endemic populations for several reasons. First of all oats would probably go relatively undamaged because of the low population levels and the tendency of endemic populations to remain in wheat. Secondly, an ecologically sound management program would involve maintenance of CLB population levels as well as suppression. [CLB densities are best maintained at just below the economic threshold (Haynes, 1974).] Making oats available to beetles as an alternative host after wheat matures in a season and is no longer attractive should serve to increase regional CLB production. This switch between crops according to regional density would thus reduce the need for insecticides and tend to maintain population levels of hosts and parasites at more stable levels. Another factor relating to planting is the planting date of the crops. By planting wheat late in the fall, its infestation can be increased. Sim- ilarily, oat damage can be reduced by planting it late in the spring. Thus planting dates can be adjusted as part of an overall regional management program although factors such as Hessian fly damage and yield reduction from late planting must be considered. Resistant wheat has an uncertain future in CLB pest management in Michigan. The current understanding is that over a period of years CLB populations are maintained on wheat and occasional population surpluses move to oats, thereby precipitating a rapid regional density increase. During phases of declining regional densities, the oat population dimin- (E sire I'C'S' oat: 52 ishes rapidly and most beetles are again found on wheat. The release of resistant wheat could have a large impact on the CLB ecosystem. The widespread planting of wheat with the current level of resistance could almost eliminate wheat as a source of CLB and parasite production. It is possible that resistant wheat would greatly reduce what would otherwise be low regional densities, even driving parasites to local extinction. During epidemic CLB populations, resistant wheat would further the shift of beetles to oats. Thus it appears that resistant wheat would reduce the small degree of stability that presently exists in the CLB ecosystem and its; release would be counterproductive. Wheat which was less resistant, or mixtures of resistant and susceptible wheat could be important, however, in reducing damage to wheat during epidemic populations, while minimizing the adverse effects described above. In areas where spring wheat is grown, resistance in wheat could be as advantageous in CLB management as resistant oats would be in Michigan. 811 of the control features discussed to this point pertain to plant- 109 01’ Crops and these decisions will generally be made in the fall. Between the fall and spring overwintering mortality can greatly affect the status 01’ CLB DOpulations. It appears that beetles in Michigan generally cannot survive winter temperature exposures above ground and they are not killed by temperatures encountered on the ground. It is possible, however, that a very mild winter or an extremely cold period without snow cover could affect this winter survival as could an early or late spring. Insecticide applications can be used against adults to "fine tune" the System just before oviposition. If it is determined that a particular field will have an excessive egg input, the proper level of reduction can 53 be achieved most economically by strip spraying. In light of the movement of CLB's between hosts the possibility of re-entry of adults into wheat or oats sprayed early in the season must be considered. When treatment is deferred until after the period of rapid movement, this problem is mini- mized, however the impact on subsequent oviposition is reduced. REFERENCES CITED Anonymous. 1974. Michigan County Statistics, Field Crops 1959-1972. Mich. Dept. Agric., Lansing, Mich. 108 pp. Asahina, E. 1969. Frost resistance in insects. Advances in Insect Physiology. 6: 1-49. Atwal, A. S. 1960. Influence of temperature and duration of conditioning on oxygen consumption and specific gravity of the haemolymph of Anagasta (Ephestia) kuehniella (Zell.). Can. Jour. Zool. 38: Castro, T. R. 1964. Natural history of the cereal leaf beetle, Oulema melano a (L.), and its behavior under controlled environmental con tions. Ph.D. Thesis, Michigan State University. 121 pp. Castro, T. R., R. F. Ruppel, and M. S. Gomulinski. 1965. Natural history of the cereal leaf beetle in Michigan. 0. Bull., Michigan State University Agric. Exp. Stn. 47: 623-653. Chiang, H. C., D. Benoit, and J. Maki. 1962. Tolerance of adult Drosophila melanogaster to sub-freezing temperatures. Can. Entomol. 94: 722-7. Curl, L. F. and R. W. White. 1952. The pink bollworm. Ip_Insects. The Yearbook of Agriculture. U.S.D.A., U.S. Gov. Printing Office, pp. 505-510. Denton, W. H. 1973. Overwintering in the cereal leaf beetle, Oulema melano us (L.) (Coleoptera: Chrysomelidae). Ph.D. Thesis, Purdue University. 140 pp. Gage, S. H. 1974. Ecological investigations on the cereal leaf beetle, Oulema melano us (L.), and the principal larval parasite, TetrastiChus jfilis (Walker). Ph.D. Thesis, Michigan State University. 172 pp. Gage, S. H. and D. L. Haynes. 1975. Emergence under natural and manipu— 1ated conditions of Tetrastichus julis, an introduced larval parasite of the cereal leaf beetle, with reference to regional population management. Environ. Entomol. 4: 425-434. 54 55 Green, G. W. 1962. Low winter temperatures and the European pine shoot moth, Rhyacionia buoliana (Schiff.) in Ontario. Can. Entomol. 94: 314136. Greenbank, D. O. 1970. Climate and the ecology of the balsam wooly aphid. Can. Entomol. 102: 546-78. Haynes, D. L. 1973. Population management of the cereal leaf beetle. Ip_ Insects: Studies in Population Management. Geier, P.W., Clark, L. R., Anderson, 0. 0., and Nix, H. A. (Ed.). Ecol. Soc. Aust. (memoirs 1): Canberra. Helgesen, R. G. 1969. The within-generation population dynamics of the cereal leaf beetle, Oulema melanopus (L.). Ph.D. Thesis, Michigan State University. 96 pp. Helgesen, R. G. and D. L. Haynes. 1972. P0pulation dynamics of the cereal leaf beetle, Oulema melanopus (Coleoptera: Chrysomelidae): a model for age specific mortality. Can. Entomol. 104: 797-814. Krueger, H. R. and R. D. O'Brien. 1959. Metabolism and differential toxicity of malathion. Econ. Entomol. 52: 1063-7. Monroe, R. E. and C. S. Polityka. 1965. The comparative toxicities of three insecticides to the cereal leaf beetle. Mich. State University Agric. Exp. Stn. 0. Bull. 48: 140-3. O'Brien, R. D. 1961. Esterase inhibition in organophosphorus poisoning of house flies. Econ. Entomol. 54: 1161-4. Pantyukov, G. A. 1964. The effect of low temperature on different popula- tions of the brown tail moth, Euproctis chrysorrhoea (L.), and the gypsy moth, Lymantria dispar (L.). Entomol. Rev. 43: 47-55. Pielou, E. C. 1969. An introduction to mathematical ecology. Wiley- Interscience, New York. 286 pp. Raske, A. G. 1975. Cold-hardiness of first instar larvae of the forest tent caterpillar, Malacosoma disstria (Lepidoptera: Lasiocampidae). Can. Entomol. 107: 75-80. Ruesink, W. G. 1972. The integration of adult survival and dispersal into a mathematical model for the abundance of the cereal leaf beetle, gglema melanopus (L.). Ph.D. Thesis, Michigan State University. PP- Ruesink, W. G. and D. L. Haynes. 1973. Sweepnet sampling for the cereal leaf beetle, Oulema melanopus. Environ. Entomol. 2: 161-72. 56 Ruppel, R. F. and G. E. Guyer. 1972. Infestation and control of the alfalfa weevil and cereal leaf beetle in Michigan, 1969-1971. Res. Rep. 164, Farm Sci. Series, Michigan State University Agric. Exp. Stn., East Lansing, Michigan. 7 pp. Saini, M. L. and H. W. Durough. 1970. Persistence of malathion and methyl parathion when applied as ultra-low volume and emulsifiable concentrate sprays. Econ. Entomol. 63: 405-8. Salt, R. W. 1950. Time as a factor in the freezing of undercooled insects. Can. Jour. Research. 0. 28: 285-91. Salt, R. W. 1958. Application of nucleation theory to the freezing of supercooled insects. Jour. Ins. Physiol. 2: 178-88. Salt, R. W. 1961. Principles of insect cold-hardiness. Ann. Rev. Entomol. 6: 55-74. Salt, R. W. 1966. Relation between time of freezing and temperature in supercooled larvae of Cephus cinctus Nort. Can. Jour. Zool. 44: 947-52. . Schillinger, J. A., Jr. and R. L. Gallun. 1968. Leaf pubescence of wheat as a deterrent to the cereal leaf beetle, Oulema melanopus. Ann. Entomol. Soc. Amer. 61: 900-3. Somme, L. 1965a. Further observations on glycerol and cold-hardiness in insects. Can. Jour. Zool. 43: 765-70. Somme, L. 1965b. Changes in sorbitol content and supercooling points in overwintering eggs of the European red mite [(Panonychus ulmi (Koch)]. Can. Jour. Zool. 43: 765-70. Somme, L. 1967. Studies on cold-hardiness in insects. Universitetsforlaget. Aas & Wahl, Oslo, Norway. Southwood, T. R. E. 1966. Ecological methods. Methuen, London. 391 pp. Sullivan, C. R. 1965. Laboratory and field investigations on the ability of eggs of the European pine sawfly, Neodiprion sertifer (Geoffroy). to withstand low winter temperatures. Can. Entomol. 97: 978-93. Sullivan, C. R. and 0. R. Wallace. 1972. The potential northern dispersal of the gypsy moth, Porthetria diSpar (Lepidoptera: Lymantrudae). Can. Entomol. 104: 1349-55. Taksdal, G. 1967. The ecology of cold-hardiness in different populations of the black currant gall mite, Cocidophyopsis ribis. Entomol. Exp. and Appl. 10: 377-86. Tripathi, 57 R. K. and R. D. O'Brien. 1973. Effect of organophosphates in vivo upon acetylcholinesterase isozymes from housefly head and thorax. Pesticide Biochem. and Physiol. 2: 418-24. Tummala, R. L., W. G. Ruesink, and D. L. Haynes. 1975. A discrete com- ponent approach to the management of the cereal leaf beetle ecosystem. Environ. Entomol. 4: 175-86. Webster, J. A., S. H. Gage, and 0. H. Smith, Jr. 1973. Suppression of Wells, M. Wellso, S. Wellso, S. Wellso, s. Wilson, M. Yun, Y. M. the cereal leaf beetle with resistant wheat. Environ. Entomol. 2: 1089-91. T. 1967. Evaluation of methods of chemical control of the cereal leaf beetle (Oulema melanopus, L.) with respect to an integrated plan. M. S. Thesis, Michigan State University. 55 pp. G. 1973. Cereal leaf beetle larval feeding, orientation, de- velopment, and survival on four small-grain cultivars in the laboratory. Ann. Entomol. Soc. Am. 66: 1201-8. G. 1974. Aestivation in relation to oviposition initiation in the cereal leaf beetle. Ip_Chronobiology. Edited by L. E. Scheving gt_a1, Igaku Shoin Ltd.: Tokyo. pp. 597-601. 6., R. V. Connin, R. P. Hoxie, and David L. Cobb. 1970. Storage and behavior of plant and diet-fed adult cereal leaf beetle, Oulema melanopus (Coleoptera: Chrysomelidae). Mich. Entomol. 3: 102-107. C. and R. E. Shade. 1966. Survival and development of larvae of the cereal leaf beetle, Oulema melanopa (Coleoptera: Chrysomelidae), on various species of Gramineae. Ann. Entomol. Soc. Am. 59: 170-3. 1967. Effects of some physical and biological factors on the reproduction, development, survival, and behavior of the cereal leaf beetle, Oulema melanopus (L.) under laboratory conditions. Ph.D. Thesis, Michigan State University. 153 pp. APPENDIX A A COMPUTER ANALYSIS OF STRIP SPRAYING FOR THE CONTROL OF CEREAL LEAF BEETLES Introduction For the control of active arthropod pests, an alternative to blanket- ing a habitat with pesticides is the technique of applying the pesticide in selected areas where they contact it as a result of their mobility. Depending on the mobility and susceptibility of target and non-target species in temporal and spatial association, this approach can result in differential mortality between species and a reduction in the amount of pesticide applied. Adults of the cereal leaf beetle (CLB) are quite active inearly spring and when beetle densities are high they can readily be seen moving between host plants in frequent short flights in apparently random direc- tions. This mobility, coupled with a high susceptibility to insecticides suggests that a relatively persistent organic phosphate insecticide such as malathion sprayed in narrow strips in a grain field could have an impact on the CLB population throughout the field. This section describes efforts to evaluate this control strategy through computer simulation. When a field is sprayed in strips as in Fig. Al. there are several factors important in determining the mortality of its inhabitants. The first factor is the direct effect of the chemical on those arthropods in the path of the spray. After this initial effect, the movement of the 58 59 7‘7 1—___—__T UNSPRAYED SPRAYED _ UNSPRAYED 3. 3 SPRAYED 2T7? UNSPRAYED E SPRAYEo- F _ ‘2. UNSPRAYED g 3 - 5 SPRAYED 8___ UNSPRAYED ( SPRAYED Fig. Al. Two strip spray patterns showing a strip comprised of a sprayed and an unsprayed portion. (Neg. 752244-14) 60 the survivors and the residual effect of the pesticide determine the final level of mortality. In evaluating strip spraying for CLB‘s, the first of these factors is easily evaluated. The residual effects of the pesticide are quite complex however, and involve the degradation of the pesticide in the field, the movement into and out of the sprayed areas, the accumulation and degradation of the pesticide within the insects, and the effects of the pesticide on the insects. These factors were evaluated in field and laboratory studies. Their inclusion in a model allows computer simulations of experiments involving a variety of strip configurations, insecticide concentrations. and movement rates. The evaluation and modeling of many of these factors was greatly simplified by imposing some constraints on the Spray conditions. The model is intended to simulate application of malathion in narrow strips in a grain field by a tractor-drawn sprayer about 11 a.m. on sunny days when the temperature is above 60°F and the wind speed less than 10 m.p.h. Although these constraints greatly simplify the modeling, they do not greatly hinder the applicability of the results as they reflect the standard or Optimal conditions for application. Because strip spraying against CLB adults would normally be implemented in early spring before CLB para- sites emerge from overwintering the effect on parasites is not included in this analysis. Methods and Results Beetle Movement Cereal leaf beetle movement was studied by observing undisturbed in- dividual beetles in grain at Gull Lake. The rows of wheat and oats were 61 evenly spaced at 6 in. apart and the beetles were clearly visible on the small plants. On each of the dates listed in Table A1 several beetles were observed for about 5 minutes each, during which flight distances (total displacement in the x -y plane) were estimated (using the row spacing as an index) and the time between flights was measured with a stopwatch. These data were recorded on a cassette tape recorder to facilitate tracking the beetles. Table A1 summarizes the results of these observations collected over a 3-year period. Theoretical Background and Analysis A l-dimension diffusion model was used to describe beetle movement in the strip spray model. The single dimension is justified because in a field sprayed as in Fig. Al, only movement in the x direction results in beetles getting into or out of or nearer or farther from the sprayed strips. Since the field lengths are very much greater than the strip widths, movement out the ends of the field is considered negligible. A diffusion model is justified because the movement of beetles within the fields involves flights of distances very much smaller than the dimensions of the field and the time interval between flights is measured in minutes compared to a scale of days for a strip spray experiment. Thus relative to the magnitude of the temporal and spatial dimensions of a strip spray experiment, the movement of beetles approximates continuous motion. Pielou (1971) shows that for a l-dimension diffusion model without drift, a diffusion coefficient (0) can be calculated by 044113 (1) ZAt where AX is the displacement in the x direction. For a 2-dimension model 0 is calculated as 62 D=LAJQE (2) 4At where an is the displacement in the x-y plane. Using Pielou's assumptions that the x and y components of displacement (z) are independent and identically, normally distributed with mean zero and unknown variance, it can be readily shown that the D for the l- and 2-dimension models are identical, hence eq. 2 can be used to calculate D for l-dimension diffusion. Thus 0 can be calculated as the mean of the individual observations of 22/t. Since not all observations on 2 and t in Table A1 are paired, the mean 22 and mean of t-were used to calculate D for each sample date. The probability of an insect being at a distance X at time t is approximated by a normal probability density function with a mean of 0.0 and a variance of 20t. The equation 2 = 1 -x POM) 41-Dt exl”(Amt (3) uses the diffusion coefficient of eq. 1 to calculate the probability of an insect moving distance X in time t. A multiple regression analysis of D vs. the main environmental factors listed in Table A1 indicated a highly significant correlation with the time of the season (in degree days). The diffusion coefficient decreases as the season progresses. There is no significant correlation between time of day, temperature, or solar radiation and diffusion rates; however, it should be remembered that consistent with the constraints discussed earlier, these observations were made under relatively uniform conditions. (Under more extreme conditions such as temperatures below 55° and wind speeds over 10 m.p.h., CLB movement is greatly reduced). Cereal leaf beetles were observed to not move appreciably between sunset and sunrise. For purposes of simulation, a constant rate of diffusion was assumed for the day with no movement at night. 63 Table A 1. Observations on cereal leaf beetle movement, 1972-1974, in oats (O) and wheat (W) at Gull Lake. "? “2. 33 .5 4L 2 2: 2 4: (D u. n. m rs +3 +4 o a 9- 5 :5 :5 .- ”.2; 0 OJ >~, Q. - ‘U (Ut— D 0 $- 0 ‘6—\ 4: E on o E : r—I‘U v m lL m ‘l-N £3 :3 é? 23 #3 E; 6341 b< 23 b< 23 25:1 1972 May 22 3 PM 386 6" O 80 -- .85 341.32 22 5.835 14 497.90 24 4 PM 428 7" 0 82 -- .80 120.80 50 4.759 39 143 72 24 8 PM 428 7" 0 74 -- .05 749.13 18 1.771 9 331 64 25 10 AM 449 7" O 78 -- .95 151.18 55 2.830 42 106 96 June 5 3 PM 623 9" O 76 -- 1.15 59.56 32 3.026 26 45.06 6 11 AM 642 10” O 75 -- 1.10 174.56 10 4.829 9 210.69 6 1 PM 642 10" 0 77 -- 1.20 21.12 26 2.968 17 15.67 7 10 AM 657 ll" 0 72 -- .85 39.87 25 4.492 22 44.77 7 2 PM 657 11" O 77 -- 1.20 89.14 21 3.231 16 72.00 13 11 AM 746 10" 0 77 -- .20 110.90 30 4.020 24 111.45 14 3 PM 776 ll" 0 86 -- .70 401.76 13 2.738 10 275.00 1973 May 20 1 PM 366 4" O 63 4.38 1.00 521.80 15 2.213 14 288.69 20 1 PM 366 11" W 63 4.38 1.00 284.31 39 5.444 33 386.95 26 4 PM 441 15" W 66 4.66 1.30 129.52 9 1.212 6 39.24 June 7 2 PM 637 ll" 0 76 -- .80 15.68 6 ,2.396 5 9.39 7 6 PM 637 ll" 0 76 -- .80 250.79 33 4.719 30 295.87 8 11 AM 661 ll" 0 75 8.42 -- 105.18 13 3.308 10 86.98 8' 3 PM 661 ll" 0 83 8.42 -- 370.85 14 4.835 13 448.26 9 11 AM 689 12" O 78 4.74 .90 68.06 8 0.585 6 9.95 10 1 PM 713 12" O 83 4.95 1.20 48.31 11 1.497 12 18.08 12 2 PM 763 13" O 85 9.00 -- 28.90 7 0.611 7 4.41 15 1 PM - 837 14" O 75 5.35 .90 76.51 16 2.038 13 44.15 1974 June 4 11 AM 590 -- O 76 3.64 1.10 41.82 15 5.339 15 55.82 4 1 PM 590 -- 0 81 3.58 1.10 30.87 9 7.665 9 59.15 25 11 AM 985 -- O 66 7.15 .90 38.70 12 2.713 6 26.25 27 ll.AM 1023 -- O 78 5.83 1.20 58.80 12 0.994 6 14.61 64 Insecticide Features: Decay Rate and Insecticide-Induced Morta1itx One set of experiments was conducted to eva1uate the decay rate of ma1athion on cerea1 grain p1ants, the dose-morta1ity response of CLB's from exposure to residuaI ma1athion. and the morta1ity of beet1es directIy exposed to the ma1athion spray. In these tests a 1aboratory spray apparatus was used to simu1ate appiication of 1.5 1bs/acre of active ma1athion E.c. in 40 ga1. water by a tractor-drawn sprayer. Twenty adu1t cerea1 1eaf beet1es from Ga1ien, Michigan, were p1aced on 4" pots with 6" bar1ey seed1ings spaced at 1east % inch apart and the p1ants were sprayed at 11 a.m. on June 2 and June 10. 1974. Ten minutes after appIication the beet1es were removed from the p1ants and the p1ants were p1aced outdoors on grass. In both tests. the sprayed beet1es were a11 knocked down by 10 minutes, and at 24 hours morta1ity was measured to be 100%. At the time interva1s indicated in Tab1e A2. some of the treated p1ants were moved indoors to a 70° room where 20 adu1t CLB's were confined on each p1ant by means of a gIass Iantern g1obe with a screen top. With some occasiona1 encouragement to keep a few off the gIass, the beet1es remained on the p1ants for the duration of the exposures. After the various exposure times indicated in Tab1e A2, a samp1e of 40 beet1es was removed from two pots of p1ants and he1d for 24 hours before mortality was counted. The results of these tests show the re1ationship between time of exposure and mortality as the ma1athion decayed during the two days foiIowing application. It can be seen that as the insecticide decays, the exposure time re- quired to cause a given 1eve1 of mortality increases. When percent mor- ta1ity is p10tted against the 10g of exposure time for each test a series 65 HH.H@ ocm «a.mm omm mm.mm mNH oo.om omv mm.om omH om.~m om mm.m~ ecu mo.m~ om om.mm om mm.H~ oNH Hm.om om om.- om mm.m om om.~ om oo.o~ mH mpwpmucoz A.cwzv xpwrmucoz 1N.:wzw .Nuwpmugoz A.c.zv a meme . a wave a we?» mgamoaxu mcamoaxm mcamoqu oo.u~ . om.HH mN.m :o2amu.Faa< capcq .230: mama .oH uzae 5H.nm oqm m~.Hm oou oo.oo om“ ¢~.om omH mm.on om mH.Hm com na.mn om" om.mm omH mm.mm QNH om.~m oe om.mm mmu om.me omH cm.mm omfi oo.on om om.~o om mH.m~ mmfl mm.o~ ooH mm.mm om mn.om me N¢.mm om o om mm.m om om.nm om om.- om Hm.om oH auwfimpgoz h.:rzv aumeucoz a.cwzv .xuwfimucoz ~.cwzv xwwempcoz A.cwzv PNumpmuLoz A.:.zv a we.» a we?» a were a .mewh a wave mczmoaxm mczmoaxm mgzmoqu mgzmoaxm mgzmoqu mu.~m om.- om.m m~.~. mu. garbagepaa< ampc< ”Lao: mumfl .N wzan .mpcmpa cavemen corcuapms op mappmmn mmmp megmu prawn acrmoaxm co mummy 03» mo mupzmam .~ < mrnmp 66 of sigmoid graphs resu1ts and, hence. in Fig. A2 the time and morta1ity axes are expressed in 109 and probit sca1es, respective1y. Regression equations were fit to each of the data sets in Tab1e 2 (average r2 = .811) and Tab1e A3 was generated by soTVing each of these equations for percent morta1ity. Tab1e A3 serves as a standardized data set. a110wing ana1ysis of both the insecticide decay rate and the dose-morta1ity response. The morta1ity 1eve15 in Tab1es A2 and A3 are determined by the dose of insecticide that the beet1es were exposed to. In these exposures. the amount of this dose is determined by the concentration of insecticide at the time of the exposure [C(t)] and the duration of the exposure (T). The product C(t) x T is the amount of insecticide accumuTated during an exposure and is termed exposure 1eve1. ActuaTTy. the pesticide decays somewhat dur— ing an exposure period, but since the decay rate is very s1ow compared to the duration of exposures in this experiment, the concentration is assumed constant during exposures. Thus. the concentration during the time interva1 (t1. t2) is approximated by C(tl). Saini and Durough (1970) showed an exponentia1 decay of ma1athion on cotton p1ants. Assuming a simiIar decay on sma11 grains. the residua1 con- centration of ma1athion can be expressed by an equation of the type C(t) = Coe‘Kt ‘ (4) where Co = initia1 concentration = 1.5 Tbs/acre in the 1ab experiments. K = decay constant, t = hours after app1ication. Thus, the exposure 1eve1 of an insect exposed from time t1 to t2 is E1 = COG-Kt1(t2 - t1) = Coe-Kt1T (5) Where T = duration of exposure. 67 Ammueemmmn .mmzv .mbcmpa as» ou mu_uvuummcm asp mcpzpaao cuppa wee.“ mzowgm> pm zurpmysoe use mucosa vmummc» co a.mgo to we.» «Lamonxm :mozpmn ahgmcovumch mg» .~< .mvm umamomxu muth_2 000. 00¢ CON 00. 00 0e. ON 0.0 m ¢ ~ _ — - o :q_fi_—d a u—A—__—— A _—_ d— 2 C) (D I) It) (3 p. ALI‘IVLUON % CD CD mm a.mm 68 mm.nmm.m mm.mmm.~ mm.on mm.¢em ~¢.mmm om.cmm om.mnH mn.mn om om.Hom.H “a.mmm.fl mm.m~m me.~¢~ oa.eo~ mm.mmH oH.moH ¢H.m¢ om us.noo.~ mm.Hmm mn.om~ mm.uwfi em.mmfl mm.mm mm.om mm.mm on mn.o~m mm.HHm on.mmH mm.HmH mo.mm om.~m N~.Hm nw.¢m om mm.oom w¢.flm¢ oe.HmH «a.mmfi mm.om mw.m¢ No.mm om.mH om mfi.¢mm ~¢.eom Hm.mHH Hm.HoH mm.n¢ om.mm om.m~ H~.¢H oe oo.mmm om.mo~ mm.mm mm.am mH.mm om.o~ cm.om mm.m om om.mefl eo.mmH mm.om om.mo Nm.HN mm.mfl mo.¢~ em.“ om en.mn mm.~n mm.nm mm.¢¢ mm.HH Hm.oH mm.m m¢.¢ oH ugmwrzmo mesa: mn.m~ oo.mH om.mH om.oH om.m mm.m m~.¢ mm. wipmugoz >Sam Lmum< mgzoz mn.Hm oo.m~ om.N~ om.~H om.m m~.m m~.¢ mm. x ommmzomm mmamoaxm mMPDZHz .cowumupFaam muvuvpumma tween mmewp pcmgmmmwu pm m~m>mp xuwpmugoe maowgw> Low umngcmg mgamonxm mmuscwz .m < mrnmk 69 Since the exposure level of a population uniquely determines the mortality level, it follows that groups of beetles with similar mortality levels experienced similar exposure levels. Thus, for a particular mor- tality level in Table A3, for instance 50%, all the combinations of C(t) x T that cause 50% mortality must result in the same exposure level (E150). This can be expressed as: E150 = C(t) x T = coe'Kt‘i or in linear form: 1n E150 01‘: 111(COT) which may be recognized as the linear equation relating CDT to t]. Thus. 1n(CoT) - Kt1 (6) if ln(coT) is plotted against t]. the resulting straight line has a slope = K and an intercept = log(E150). Fig. A3 shows plots of ln(CoT) vs. t.l for the 10 to 90% mortality values in Table A2. In this plot. the X axis has been converted to hours daylight after spray application (using a lS-hour day) as this transfor- mation resulted in a better fit of the data to equation 4 (average r2 = .939). The results of the regression equations fit to these data are shown in Table A4. Ideally, the 9 values for K in Table A4 should all be identical since they all reflect the decay rate of the insecticide. As seen in Fig. A3 and in the table. these slopes do not differ greatly; however. because of the observed differences, the 50% mortality values were used to develop the equation expressing the residual concentration of malathion as: C(t) = coe-.1585td (7) where td = hours daylight since application. This residual concentration is shown graphically in Fig. A4. 70 . mmuwammmm .mm as?» cm mzmumu muvuwpummcw ms» nu mrw>mp aerauLos mzorcm> com uwgrzaoc mwgsmwnxm muwurpummuw .m< .mFu 295%:ng wo_o_._.ommz_ awkmd HIO_.._>oF «Lamogxm :mmzpmn ngmcovumpmg ms» .m< .mwm 3500\dn. x625 4w>m.— wmamon—xu 000. 00¢ OON 00. om 0* ON 0. 0 ¢ N _ :d—_q_. — —_.‘-——_ - —_————_. — — #0.. + :35 x 3m. a EROS: Enema 1 l l o o o o I~ In In — ALI'IVLHOW ‘16 1 O m [m.mm 75 No behavioral differences were observed between the CLB's confined to treated plants and those controls confined to untreated plants. Also. no change in behavior was observed as the insecticide exposure of the beetles increased. Without any outward indication of stress. beetles which had accumulated a high enough dosage would fall from the plants and remain on their backs kicking. To test for a possible repellent effect of the ma1athion on CLB's, 4 treated (1.5 lbs/acre) and 4 untreated plants were placed in a 2 x 2 x 2 ft. screened cage and 20 CLB's were introduced at various times after the insecticide application. At the time the first CLB fell from insecticide exposure (first knockdown), the distribution of the 20 CLB's in the cage was determined. The numbers on the sprayed and unsprayed foliage as well as the number elsewhere in the cage were recorded. The results shown in Table A5 show no significant difference at the .05 level in the numbers on the sprayed and unsprayed plants. The effect of the insecticide on activity was measured in an experi- ment where 8 treated plants (1.5 lbs malathion/acre) were placed in one 2 x 2 x 2 ft. cage and 8 untreated plants were placed in a second similar cage. One hour after the spray was applied, 20 CLB's were introduced into each cage. At 2-minute intervals. each cage was observed for 10 seconds land the number of CLB's which moved during that time was recorded. 0b- servations were ended with the first CLB knockdown at 25 minutes. The results shown in Table A6 indicate no difference in activity between the CLB's in the 2 cages. Although the experiment was terminated at 25 minLJtes with the first knockdown. the beetles' exposure was fairly large. According to the insecticide model of eqs. 7 and 8, the exposure they Table A 5. Distribution of 20 cereal 1eaf beet1es in a cage at the 76 time of first knockdown. First No. on No. on No. Minutes After Knockdown Sprayed Unsprayed Elsewhere Application (Min.) Plants Plants in Cage 15 5 9 6 5 30 25 7 13 60 25 9 11 0 90 30 7 7 6 7' 8.00 9.25 2.25 S.D. 1.15 3.30 3.20 77 Table A 6. The number of cereal 1eaf beet1es moving out of 20 during a 10 second observation period at various times after intro- duction into cages of sprayed and unsprayed p1ants. Time After No. Mov1ng in 10 Sec. Introduction Sprayed Unsprayed 2 minutes 1 3 4 ll 6 ll 0 H 8 u 10 " 12 " 14 " 16 " 18 " 20 " 22 " NU'INWWNNNNOJN wwmwwmbww 24 " ><| 2.583 2.750 5.0. 1.240 1.288 78 received should have ultimately resulted in 55.3% mortality. These two experiments. while somewhat preliminary in nature, indicate that malathion applied at the rate of 1.5 lbs/acre does not affect CLB behavior and, thus, behavioral effects were not incorporated into the strip spray model. Temporal and Regional Changes in Insecticide Tolerance An important factor in determining the applicability of a strip spray model is the uniformity of the insect response to the insecticide. Re- gional differences in tolerance levels, the development of insecticide resistance, and within season changes due to beetle age must be considered before widespread utilization of the technique is accomplished. Monroe and Polityka (1965) collected adu1t cerea1 1eaf beet1es near Galien. Michigan, in the spring of 1965 and gave them topical applications of malathion in l microliter of acetone. Using a range of insecticide concentrations varying from .01 to .04 micrograms/microliter, they deter- mined the L050 to be .0l8 ug./beetle. In 1973 and 1974. adult cereal leaf beet1es were collected at two sites near Galien and also at Gull Lake, 77 miles N.E., and Laingsburg, Michigan, 135 miles N.E. of Galien, in order to measure possible differ- ences from the baseline established by Monroe. The results shown in Table A7 were fit with log-probit regression equations which were solved for 50% mortality. The resulting L050 values indicate that the beetles collected in the spring of 1974 from site 1 (3 miles southeast of Monroe's collection site) had about the same tolerance level (L050 = .0184) as Monroe measured 9 years earlier. The newly emerged summer adults col- lected at the same site in July of 1973 and 1974 and similarly tested, were found to have L050's of .0444 and .0442. respectively. The 2.5-fold m: mNNo. u omoH H o Hocpcou e m.Nm oeo. c a.me one. a c.0N ONo. N N.mH mHo. c m.N oHo. mmumuwHamm NHH—mucoz amnw x mmoo emmH .mN >43a man .mH >Han H wka zmH4<¢ H mka zmHHmecH oer pcmcmmmwv cur: m.mHu op mvHquummcH mo mmmou pcmHm>H=am N mcwaHaam seem mcHqummc zuHHmucoz .23: zo.._. Fmoz Jomkzoog mmoo HEN 20mm Fmoz Elm 4 00. «N5 93 m E .Eoz 45.2.2 x .e 9.3 S. 1, 86 in time, despite the increase of mortality in the controls. This clearly demonstrates that recovery occurred, although the slope of this line is only significantly different from 0 at the .1 level. It is difficult to quantify the recovery rate from this experiment, especially since the population tolerance to malathion apparently de- creased during the test. A dose of .01 pg./beet1e which caused 1.3% mor- tality to previously unexposed beet1es on the first day of the test was found to cause 11.1% mortality to surviving previously unexposed beetles on the 8th day. This apparent loss of tolerance means that by the 8th day the beetles had recovered from the first exposure because the mortality from the second dose (about 10% from the regression line) is less than the 11.1% morta1ity caused by a single dose on the same day. This recovery rate is very much slower than the degradation of the insecticide in the field which, according to equation 6, is over 90% com- plete in one day and 99% complete in two days. It appears that recovery is not very significant in strip spraying with malathion for CLB's and this factor was not included in the model. Strip Spray Model A Fortran program was developed to simulate strip spraying with mala- thion against the cereal leaf beetle. The program (Appendix D) simulates a single strip which consists of a sprayed part and an unsprayed part as shown in Fig. Al. It is assumed that a field would contain a large number of these strips so that field boundaries would not be a significant factor. In the simulations beet1es moving out of one side of the strip were allowed to move back into the strip from the other side (toroidal symmetry). 87 Equation 3 was used to calculate the percent of beetles moving various distances in a unit of time for the particular diffusion coeffi- cient used in the simulation. Consistent with the observation that all beet1es actually sprayed with malathion at a rate of 1.5 lbs/acre are killed, an initial density of 0 beet1es/ft2 was assumed in the sprayed part of the strip. A uniform distribution of any density was assumed as an initial condition for the unsprayed strip since there are no density dependent factors in the model. CLB's were characterized by two param- eters: their location, and their exposure level. Initially, all beet1es had an exposure level of 0.0; however, whenever beet1es moved into the sprayed strips they increased in exposure level according to the duration of the exposure and the concentration of the insecticide from eq. 7. Beetles outside of sprayed strips did not "recover" or reduce their ex— posure level. In the simulations, beetles moved and the insecticide decayed only during daylight hours. During the night those beet1es on sprayed strips continued to increase their exposure levels throughout the night. At the end of the simulation, the number of beetles of all exposure levels were totalled and percent mortality was determined from eq. 8. Field Validation A field experiment was conducted at Gull Lake during June of 1973 to provide a data set for validating the model. An 11" high oat field measuring 88 ft. by 1250 ft. was divided into four stips (20 x 1250 ft.). 0f each strip, 8.5 ft. was sprayed at a rate of 1.5 lbs/acre at 11 a.m. on June 7. At various times after insecticide application, the treated field (designated as field 1) was sampled with a sweepnet and the density of 88 CLB's/ft2 was calculated using the model of Ruesink and Haynes (1973). Two control fields (designated as fields 2 and 3) were also sampled at the same times. Both control fields had about the same dimensions as the sprayed field and both had 11'I high oats on June 7. Field 2 was located 100 ft. west of the sprayed field, and was separated from it by a freshly plowed field. Field 3 was 1875 ft. east of field 1, separated by a road and several fields of alfalfa and grains. The results of these samples (Table A10) show that the density in the sprayed field (#1) declined dramatically during the first 24 hours after spraying, and then began to increase slightly. The densities in the two control fields (#2 and #3) appeared to fluctuate considerably be- tween samples. What is more probable is that the actual densities in these fields remained approximately constant and the apparent differences in density reflect the inability of the sweepnet model to adequately account for changing environmental conditions. This model was not intended for use at 7 p.m. or 9 a.m. (when the 8 and 94 hour counts were taken). It is important to note that the density in field 2 did not decrease with respect to field 3. This indicates that relatively few CLB's moved the 100 ft. from field 2 into field 1 where they would likely have been killed. Because there was apparently no significant movement of CLB's among the 3 fields in this experiment, the control fields were used to evaluate the density reduction in the sprayed field. For each of the samples in Table A10, the relative density in the strip sprayed field was determined by dividing the strip sprayed density by the average density of the two controls. Since the initial density in the strip sprayed field was 1.2% greater than the average of the two controls at the same time, each of the relative densities for this field were multiplied by .98814 to determine the 89 Table A 10. Results of a field test of strip spraying. Field No. No. % of Initial Time No. Sweeps Caught No./Acre Density 1.5 hours before spraying 1 1,500 151 1,441.8 100.00 2 700 89 1,825.2 3 700 50 1,023.7 8 hours after spraying 1 2,700 72 389.8 11.76 2 350 95 3,545.8 3 350 81 3,023.1 24 hours after spraying 1 2,700 4 27.2 2.08 2 350 23 1,058.5 3 350 32 1,472.3 49 hours after spraying 1 2,700 17 79.6 6.03 2 350 42 1,463.6 3 350 33 1,145.6 75 hours after spraying 1 2,700 15 66.1 7.92 2 350 21 705.7 3 350 28 940.9 94 hours after spraying 1 2,700 32 141.5 7.02 2 350 58 2,016.8 3 350 56 1,951.5 90 percent of the initial density. These results, included in Table A10, are corrected for sampling error and for the natural morta1ity which occurred during the 4 days of the test. The same results are plotted in Fig. A7. At several times during the strip spray experiment of June 7-10, 1973, observations were made in the nearby control field on the rate of movement of the cereal leaf beet1es. From these observations which are included in Table A1, a weighted mean diffusion coefficient of 207.46 in2/min. was calculated. Using this diffusion coefficient, the strip spray model was run to simulate the field test and the results are graphed in Fig. A7. The close agreement between the model predictions and the field observa- tions serves as a reasonable validation of the model. The slight increase in densities in the sprayed field observed in the last two days of the test might be due to the immigration of some beet1es, although no emigra- tion was detectable from the nearest grain field (#2). Model Analysis: Sensitivity Simulations were run with the model to investigate its sensitivity to the diffusion rate of the beetles, the strip configuration, and the initial insecticide concentration. All the simulations resulted in either 100% mortality or a stable morta1ity level after the insecticide completely degraded. These final mortality levels were used to evaluate the effects of changing the three parameters. The diffusion coefficients calculated from the field data on flight frequencies and distances have a wide range of values (4.41 to 497.90 in2/min.). In Fig. A8 it is seen that varying 0 over this range has a large impact on the final mortality level, however the model is less sensi- 91 Hmm-eeNNmN .mazv .mtum\.maH m.H co want a be :o_;»apas 56.3 masapm .bc om mo .uw m.m mcpzmcam m=H>Ho>cH cowbaHzeHm gouaasou a use acmewgmaxm uHmHm a mo comwcquou < .N< .mrm 20_._. com Hm>mH NuHHmucoe Hacwm mchHzmmL as» can chpm .pe ow a we umxacam sucHz as» cmmzpmn gHgmcoHumHmc mg» ucwzosm mcoHuaHaeHm gmuzasou eo muHammm AF... 0¢ u a.m.—.mv om> com Hw>mH zuHHmpsoe Hmcrm mcHHHammg me“ new awgum .ue ow a mo umzmcqm spur: esp cmmzpma awgmcoHHung mg» acrzogm mcovumHaeHm sausanu yo muHammm .m< .mwm O N 0 v 0 CD ALI'IVLHOW X 0 (I) 00. 95 costs are extremely dynamic, it is difficult to accurately assess these advantages with any degree of generality or permanence. Furthermore, in spite of a reasonable field validation of the model, there is a certain element of uncertainty associated with its predictions and this is diffi- cult to evaluate in terms of economics. For the purpose of evaluating one strategy relative to another, however, it seems that the model should be quite effective and for that purpose using relative costs for material and application should be adequate. In evaluating the economic aspects of strip spraying, it was assumed that the costs of the insecticide equaled the application cost. Using a 1975 retail price of $2.75/1b. for single container lots of malathion and an advertised cost of $2.75/acre for custom application, a cost of $5.50/acre was used for applying 1 lb. of ma1athion/acre by a tractor-drawn sprayer. In Table All, these costs were used to evaluate three strategies, all of which involve spraying 25% of the field at a rate of 1 lb/acre for a total application of 4 oz/acre. It is seen that increasing the total strip width from 20 to 80 ft. results in a decrease of 22% mortality and, thus, repre- sents a less efficient use of the insecticide. However, by utilizing the full capacity of the 20-ft. spray boom, a 75% reduction in application time is achieved in spraying the 80-ft. strip and as a result total cost of appli- cation and material is reduced by 60%. Thus, for short-term economic advan- tage, it is most efficient to utilize the 20-ft. capability of the sprayer. In Table A12, various strip widths are simulated of which 20 ft. were sprayed in each. These patterns represent the most economically advanta- geous use of the insecticide to achieve a given mortality level. All of the strategies represent a considerable reduction in both cost and material applied from the recommended application of l 1b./acre. 96 Table A 11. An evaluation of 3 strategies which involve spraying 25% of a field. Application Rate Diffusion Coefficient l lb/a re 125 in /min. Sprayed Unsprayed % Mortality Cost ozgjacre 5' 15' 100 3.44 4 10' 30' 89 2.06 4 20' 60' 78 1.38 4 97 Table A 12. An evaluation of 5 strip configurations in which 20 ft. widths are sprayed. Application Rate Diffusion Coefficient 1 lb/a re 125 in /min. Sprayed Unsprayed % Mortalityg Cost“ oz.[acre 20' 20' 100 2.75 8 20' 40' 90 1.83 5.3 20' 60' 78 1.38 4 20' 80' 66 1.10 3.2 20' 100' 59 .92 2.7 98 Discussion It is apparent from the computer simulations and field test that strip spraying can be effective in reducing populations of adult cereal leaf beetle in grain fields. The technique is not without limitations, possibly the most severe of which is deciding how much to reduce adult densities (if at all) to get the optimal egg input in a field. Once that decision is made and the approximate time decided upon, it is necessary to have two days of satisfactory weather to effectively use strip spraying with malathion. Clearly, the technique of strip spraying for CLB control needs addi- tional investigation before it can be generally implemented. Despite an encouraging field validation it is probable that the computer simulations are more valuable in choosing between alternative spray strategies than in accurately predicting a final level of mortality. Much of the work on re- covery, behavior, insecticide degradation, and beetle movement is rather preliminary and could greatly benefit from additional experimentation. The present results, however, are quite adequate to design field experiments, and leave no doubt that strip spraying can be an economical means of con- trolling CLB's. There are many other pest ecosystems where the technique and this model are of value, the primary requisites being mobility of pests and susceptibility to insecticides. Strip spraying can be advantageous in these cases in (l) effecting a differential kill of favorable and unfavor- able species; (2) reducing the amount of pesticide applied; (3) reducing the cost of control; and (4) achieving a predetermined level of control. APPENDIX B A PREDICTIVE MODEL FOR CEREAL LEAF BEETLE MORTALITY FROM SUB-FREEZING TEMPERATURES Introduction Insects can survive exposure to temperatures below their freezing points by either supercooling or by tolerating freezing (Asahina, 1969). According to Salt (1961), the majority of insects are freezing susceptible, meaning they can supercool to some extent; however, with extreme or pro- longed cold exposure they freeze and die. The remainder are able to survive freezing and are called freezing resistant or freezing tolerant. The cereal leaf beetle (CLB) is among the former group. Salt (1950) showed that freezing of insects depends not only on the temperature, but also on the duration of cold exposure. Nonetheless, the majority of published low temperature survival studies since that time have been based on supercooling point determinations. Most researchers have specified cooling rates and have attempted to use supercooling points as an index of cold tolerance. Preconditioning Many authors (Green, 1962; Pantyukov, 1964; Sullivan, 1965; Sullivan and Wallace, 1972) have observed a change in mean supercooling points through- out the winter attributable to preconditioning or cold hardening at moder— lately cold temperatures (near 32°F). Somme (1967) in summarizing some of 99 100 these results noted that diapausing insects can cold harden at relatively moderate temperatures and do not lose that hardening until after diapause termination. Wellso (1974) found cereal leaf beetles to estivate from mid-July to mid-November in Michigan. Dickler (unpublished) collected cerea1 leaf beetles in the field in July and held them at 38°F throughout the winter. He found the mean supercooling point of these beetles to decrease from 4°F in July to -10°F in October. It did not change significantly from October to February (-9.7°F), but it again increased to -3°F in April. Several authors (Atwal, 1960; Green, 1962; Sullivan, 1965; Greenbank, 1970; Sullivan and Wallace, 1972) have investigated the effects of short- term exposures at low temperatures on the supercooling point of insect populations and have found these treatments to generally lower mean super- cooling points. Three investigations have compared these acute precondi- tioning treatments to long-term, gradual preconditioning treatments and found differing results. Sullivan and Wallace (1972), working with gypsy moths (Lymantria dispar, L.), found acute exposures of 7 days at ~13°C to -22°C (8.6 to -7.6°F) to significantly lower the mean supercooling point below that resulting from 18-25 days conditioning at 0°C. However, when compared to long-term conditioning of 59-74 days at 0°C, this acute conditioning did not significantly lower the mean supercooling point. Working with the European pine sawfly [Neodiprion sertifer (Geoffroy)], Sullivan (1965) found 1 week exposures at -13 to -23°c (8.6 to -9.4°F) significantly reduced supercooling points beyond the level resulting from 8 weeks conditioning at 0°C. Similarly, Green (1962), working with the European pine shoot moth [Rhyacionia bouliana (Schiff)]. found a l-week 101 exposure to temperatures lower than 7.5°F to cause a significant decrease in the mean supercooling points beyond that caused by two month's storage at 32°F. Regional Differences in Supercooling Ability Many investigators have been concerned with the gradual development of an increased level of cold tolerance in insect populations which could result in greater overwintering survival or an increased range for an insect. To date, this phenomenon has not been investigated directly, but by comparing supercooling points of populations in different regions. Several authors (Somme, 1965a,b; Taksdal, 1967; Pantyukov, 1964; Sullivan, 1965) have reported regional differences in cold tolerance where the more cold-hardy populations were from the area with the colder winter. Other authors (Sullivan and Wallace, 1972; Green 1962) have found insignificant differences in supercooling points between populations from regions differing in cold exposures. In evaluating regional differences in cold tolerance, complicating factors such as differences in preconditioning, food quality, quantity, and contaminants such as dust particles must be considered as they influence supercooling points (Salt, 1958). There is no good evidence of regional differences in cold tolerance of the cereal leaf beetle. Dickler (unpublished) showed no consistent differences among populations collected at 3 different sites on a N-S gradient through lower Michigan. Logan (unpublished) found cereal leaf beetles from northern Michigan (Petoskey) to be consistently but insigni- ficantly more cold tolerant than beet1es from southern Michigan (Galien). 102 Predictive Models for Cold-Induced Mortality There are several examples of predictive models for winter mortality in the literature. Greenbank (1970) developed a model for winter mortality of the balsam woolly aphid by placing minimum thermometers on tree trunks next to groups of aphids. At the end of the season he correlated percent mortality with the minimum temperature of the season. This simple linear regression model described with reasonable accuracy (r2 = .72) the rela- tionship between minimum temperature and winter mortality. Green (1962) and Sullivan (1965) developed models to predict mortality of European pine shoot moth larvae and the European pine sawfly, respec- tively. Both models were based on supercooling point determinations and consisted of regressions of cumulative percent mortality on decreasing temperatures. Allowances were made in both models for changes in cold- hardiness throughout the season and both gave remarkably accurate predictions of mortality based on the lowest winter temperature and the extent of cold- hardening of the insects at that time. Since all 3 of these insects overwinter in highly exposed locations (on the bark of firs, in new pine needles, and in buds, respectively), it seems likely that their cold exposure closely approximates that of the air temperature both in extent and duration. Thus, for relatively extreme ex- posures of relatively short duration, a model based on supercooling points may be adequate. Neither of the latter two models account for the mortali- ty of those insects beneath snow cover, however, and it is questionable whether a supercooling point model could be accurate in predicting mor- tality from the milder and longer exposures in protected habitats. 103 Several authors have considered the aspects of time and temperature in developing models for mortality from cold exposure. Raske (1975) gave forest tent caterpillar larvae exposures to 6 temperatures of 3 durations and de- veloped multiple regression models expressing survival as a function of time and temperature. The models he presented for pharate larvae (fully devel- oped lst instars within eggs) and for unfed lst instars gave r2 values of .721 and .884, respectively. He used these models to conclude that low temperature exposures in the field cannot account for the high spring mor- tality occasionally observed in Alberta. Salt (1961) felt that under ideal conditions, freezing of insects is a probabilistic event and, hence, the same percent of a surviving popula- tion should freeze in one time unit as another. Thus, the compound inter- est formula Y(t) = Y(o)e"rt should describe survival at one temperature as a function of time. A consistent departure from this model (Salt, 1950, 1966) led to a conclusion that differences among individuals was a signi- ficant factor in determining mortality. Chaing gt_al, (1962) held adult Drosophila at various temperatures for exposures of various durations and found a sigmoid relationship between mortality and exposure time at a constant temperature. They pointed out the similarity of this response to that of insects exposed to insecticides, yet found their results at variance with those of Salt (1950) who used a logarithmic curve to describe the same response. Actually, in a subsequent paper, Salt (1958) acknowledged a sigmoid relationship between percent mortality and the logarithm of time and he indicated that a log-probit transformation (common to insecticide tests) linearized this relationship. 104 Salt (1966) suggested developing a model for mortality due to chronic cold exposures by holding insects at various low temperatures and determin- ing mortality as a function of time. Field temperatures are, of course, not constant so Salt suggested dividing daily exposures into a chronological sequence of constant temperature exposures. During each of these exposure periods, mortality is determined from the time-morta1ity curve for the temperature during the period. This algorithm requires time-mortality curves for many temperatures to reduce the errors due to approximating field temperatures. Salt approached this problem by determining the relationship between temperature and time required for 50% mortality (L150). By plotting the logarithms of LT50 values vs. temperatures, he developed a linear relationship between these variables allowing the continuous expression of LTSO as a function of tem- perature. Salt did not complete this model by including morta1ity levels other than 50%, nor did he discuss the validity of his implicit assumption of additivity of sequential exposures. He did clearly show a method for de- veloping a predictive model for mortality due to exposures of various tem- peratures and durations and he further showed that supercooling points are readily derived from such a model. Methods Temperature Control In order to study the effects of long-term exposures at various tem- peratures on cereal leaf beetle adults, a series of 8 constant temperature water baths were set up in a walk-in freezer using aquarium heaters and .105 motorized stirrers for temperature control. Standard lO-quart plastic pails were placed in 15" high round trash cans and the air space was filled with vermiculite. The pails were filled with a mixture of water and ethyl- ene glycol and were fitted with plexiglass covers with holes for heaters, stirrers, thermocouples, and tubes of beetles. A series of baths was set up at temperatures ranging from -5°F to 30°F at 5-degree intervals using aquarium heaters of 25 watts for temperatures of 20°F and less and 50 watts for temperatures over 20°F. Temperatures were monitored by placing a thermocouple from a recording potentiometer in each bath. Temperatures within the baths were reasonably constant despite the fact that the air temperature in the freezer fluctuated in a 8°F cycle every 20 minutes and defrosted once/day. During a typical 48-hour period the average maximum fluctuation for all baths was i .52°F (range = i .350 to i .75°F). Since the maximum deviations occurred during the defrost cycle, a more informative statistic on the consistency of temperature con- trol is the standard deviation from the mean temperature. These standard deviations averaged .360 among the 8 baths during a lZ-hour period (range = .270 to .47°F). With rapid stirring, the temperature distribution within the baths was constant to within i .l°F at all depths below 1" from the surface. The mean temperatures within the baths were found to drift in time (generally decreasing) and whenever a perceptible drift from the desired mean was observed (about .2°F) the heater control was adjusted. The aver- age time between such adjustments during a 4-month period was 22.4 days (50 = 14.1). 106 Cereal Leaf Beetle Treatment The cereal 1eaf beetles used in these tests were collected when newly emerged in early July near Galien, Michigan. They were fed in the lab on barley seed1ings until estivation, then stored at 40°F (100% R.H.). Beetles taken from storage were warmed to 70°F for about 2 hours during preparation and then refrigerated at 40°F again for an average of 3 days until tests were run. Twenty CLB's were individually placed in 12-inch long 9 mm. glass tubes where they were separated from the glass and from each other by a capsule of filter paper. Each tube was furnished with a few drops of water at the bottom before sealing the bottom with a cork and the top with a cotton plug. In setting up the tests, tubes were dropped in the baths and periodi- cally samples of 3 tubes (60 CLB's) were removed and held at 70°F for 24 hours before counting mortality. The cooling rate of the beetles in the tubes was very rapid (about 40°/min.) and, hence, the cooling period is considered a negligible part of the exposure periods which varied from 1 hour to 15 weeks. In counting mortality, those beet1es incapable of co- ordinated movement when on their feet were included among the dead. Results Mortality From Constant Exposures--Pre1iminary Model Development Table Bl lists the results of a test set up on Jan. 31, 1974* which served as a data set for the development of a preliminary model. These data show that as temperature decreases below 25°F, the time required for *The 459 results are from a test conducted on December 17, 1974. 107 Table 81. Mortality levels resulting from exposures of various temperatures and durations. -5°F 0° 5° 10° % %l % % .Hgg5§_ Mortality ng§§_ Mortality H93:§_ Mortality ng§§_ Mortality 1 18.3 1 11.9 1 1.7 1 3.4 2 27.1 6 8.3 20 18.3 100 21.7 3 55.0 10 15.0 32 31.7 174 49.2 4 67.5 14 56.7 48 21.7 240 61.7 5 83.7 15 44.1 62 33.9 308 87.9 18 60.0 88 65.0 360 94.9 22 60.0 112 48.3 431 100 26 78.3 120 100 32 89.0 140 81.7 40 100 160 88.3 188 100 15°F 20° 25° 30° %: %’ % % Hours. Mortality H92§§_ Mortality Hours. Mortality ngy§_ Mortality 240 5.0 360 12.1 360 11.7 120 3.3 360 23.3 744 25.0 744 20.3 360 28.3 480 42.4 960 34.1 960 28.8 744 10.0 744 75.0 1320 81.4 1320 37.3 960 30.0 984 98.3 1800 91.3 1800 73.3 1320 30.0 2160 100 2160 59.3 1800 72.9 2496 80.0 2160 72.9 2496 88.3 108 mortality greatly decreases. When percent mortality is plotted against time at a temperature, a sigmoid curve results and, hence, probit regres- sion lines were fit to each of the data sets in Table Bl (average r2 = .934, SD = .0508). The resulting curves (Fig. B1) were found to have an average intercept of 3.8% (range = .3 to 8.2%) mortality from 0 hours ex- posure. This initial mortality was verified by measuring the mortality due to handling alone without any cold exposure, and it was found that of 180 beetles checked just prior to exposure, 6 had already died (3.3%). To develop the relationship between temperatures and mortality rates, the equations fit to the data in Table Bl were solved for the number of hours required for mortality levels of 10 to 90%. The resulting values (Table 82) provided a standardized method of comparing temperature treat- ments. It can be seen in both Fig. Bl and Table 82 that the results at 25°F and 30°F are quite similar and beetles do not survive longer at 30° than at 25°. There is, however, a large difference in survival times between 20° and 25°. This may be interpreted as demonstrating that 25°F is near the upper threshold for low temperature-induced mortality. In Fig. 82, the times required for 50% mortality (LT50) values from Table 82 are plotted against temperatures on the upper X-axis and degrees below 25°F on the lower X-axis. This transformation of 25-T allows the development of a model based on degree hours below a somewhat arbitrary threshold of 25°F. A linearized transformation of this graph is shown in Fig. 83 where the logarithm of the Y-axis is plotted against the X-axis squared. A linear regression equation Y = 7.088 - .0067425X was fit to the transformed points in Fig. B3 with an r2 of .996. This equation gives 109 He~-HHN~mH .uazv .mmczueemaemu mzowee> o» a.mHu mo mmeamoaxm maochucou seem mchHamme mHm>mH HHHHmpeoz mmnmoaxm manor 000m 000N 00NN 000. 00+». 000. 000 CON 0 1 d H H H d H H H H H H H H H\ .J \ H \\\\.|\\\\\ 1 L ‘ L omN eon o0N on. .0. on .0 \\ \ Hm .mHa O N <3 <3 «3 1: ALIWVLBOW % O O 00. 110 Table 82. Hours exposure required for various mortality levels at various temperatures. % Temperature (°F) Mortality -5 O 5 10 15 20 25 30 10 .33 2.92 13.47 53.02 286.86 365.61 220.36 366.01 20 1.24 7.85 39.22 99.48 375.49 596.42 691.16 753.97 30 1.90 11.43 57.94 133.28 439.94 764.28 1033.57 1036.13 40 2.46 14.46 73.74 161.79 494.33 905.91 1322.47 1274.21 50 2.98 17.26 88.37 188.19 544.68 1037.06 1589.97 1497.64 60 3.50 20.06 102.99 214.59 595.03 1168.20 1857.47 1715.08 70 4.05 23.08 118.79 243.11 649.42 1309.83 2146.38 1953.15 80 4.72 26.67 137.51 ,276.90 713.87 1477.69 2499.48 2235.31 90 5.62 31.60 163.26 323.36 802.50 1708.50 2959.59 2623.28 111 AHneeNNmm .mmzv .mmcapeemnEmH moneo> pa HomHHV HHHHaHsoe Non gee umgwscmc «Lamonxm mesa: 23.3. m- o m o. 9 ON 8 on .._. $555..sz 0.» mu 8 m... o._ m m m- o m. o . n U S .09. I. nu nu 1 n: Av 88% w" nu z w o v n Hoomfl . m o .N O . Hunxw_ .8 .22 112 HNueeNNmN .mmzv . Hmsapacmaemp . omNV oH umpew>cou mH memux as» use mHeom uHssuHeemoH a we; meelz as» mean: mezmmemQEou use oth comzumn nwgmcoHpeHme umNHemmcHH < .mm .mve nxxz Axym nxgw nxyn .200 .900 “xv? .90» HXQN nxU. l 039.000.1005.; ca: (“11) unuaow egos emsnvo aansoa xa sunow 113 all the time-temperature combinations resulting in 50% mortality. The slope of this line shows the rate of loss of effectiveness of increasing temperatures in causing mortality of the CLB and, thus, gives the relation- ship needed to combine time and temperature into a standardized dose of cold exposure (called exposure level). Exposure level is thus calculated by the equation: 51 = t49.0067425 T2 (1) where E1 = exposure level; t = hours exposure; and T = 250 - temperature (in °F). Given equation 1 relating exposure level to time and temperature, it is necessary to develop the relationship between exposure level and mor- tality. This was accomplished in Table 83 by solving equation 1 for the time-temperature combinations in Table 82. The average exposure levels corresponding to mortality levels of 10 to 90% in Table 83 were plotted in Fig. 84 showing a sigmoid relationship between exposure level and mortality. By transforming the Y-axis to probits, the following linear regression equation was fit to the transformed points (r2 = .998): M = .001376 El + 3.328 (2) where M = probits morta1ity; E1 = exposure level. The curve in Fig. 84 results from solving equation 2 for percent mortality. Equations 1 and 2 together form a model for predicting mortality due to constant exposures of various durations at various temperatures. To determine how well this model fit the original data set on which it was based, the results in Table 81 were compared to the model predic- tions for similar treatments. A linear regression analysis of the observed and predicted mortalities gave a slope of .963 and an r2 of .859. This slope is very close to the l-to-l ratio of a perfect prediction, and the 114 mm.mHm mH.¢¢HN mm.mmmN mo.oHON om.NNmH cH.¢m¢H Ho.NNvN HH.anN ov.NN¢N om mo.wH¢ mn.onH m¢.mm¢N um.m¢mH mH.mmMH em.NmNH oo.c¢ON mm.momH Nn.wm0N ow um.H¢m MN.NmmH mm.ovHN om.mva mm.NNNH om.moHH mN.NmNH om.ommH Nm.m¢mH om mn.omN me.mmmH N¢.nmmH m¢.man oN.mcHH mN.mnm mw.~NmH mm.ommH mN.HHmH om mo.mNN om.¢HNH mm.mmmH mN.MNNH mo.nmcH mm.mmm mm.onH mN.noHH mH.NmNH cm mH.mNH mH.mmoH mv.NNmH “a.mooH mm.mmm nm.mmn mm.mmoH mm.mmm mm.NooH ow mn.mNH mm.omm mm.mmoH mw.Hom mN.Nmm om.noo om.¢mm Ho.mmn no.0Nm om om.moH mm.vom 0H.Hmm N¢.mom mm.mmn Hm.mm< ¢N.Hmm om.omm oo.mmm ON Nm.NmH mo.mmN om.ONN N¢.qu mN.Nom HN.H¢N mw.mmH w¢.NmH em.N¢H oH .o.m .M mN 0N mH oH m o mu HHHHMNLOZ Huey mezpmemnsm» .Nm mHne» cw mpcmEHmmeu mezumcmaemulmewp we» ou mcwucoammecoo mHm>mH mezmoaxm .mm mHnm» 115 AFFlcemNmn .cmzv .XHVPMHLOE 0:0 P0>0_. Esmoaxm cmmzumn Qwsmcowvmpm... of. .Vm .0:— HmomN v manor wwmomoju>w4 mmamomxw oooN 000N 000. 000. 00m. 000m ON 0.... ALITIVLHOW °lo 116 r2 indicates that the model explains 85.9% of the variation associated with all the observations in Table 81. The model is more accurate at the low temperatures and a similar regression of observed and predicted mor- tality levels for -5° and 0° gives r2 values of .985 and .932, respectively. If the overall coefficient of determination (r2) of .859 is compared to the average r2 of .934 for the individual probit lines in Fig. 81, it can be seen that the inclusion of temperature effects in the model resulted in a loss of .075 in the r2. Seasonal Change in Cold Tolerance As previously discussed, Dickler (unpublished) found the cold tolerance of adult cereal leaf beetles remained relatively constant from October to February, but decreased considerably by April. To further investigate this phenomenon, adult CLB's were given cold exposures (as described in the previous section) at various times during the winter. The results of these exposures are presented in Table 84.along with predicted results from the model described previously. The LT50 values were calculated by fitting a probit regression line to the results of each test and solving the equations for the exposure time causing 50% mortality. It is apparent from Table 84 that there are no large or consistent differences between the mid-December and the early February tests. By mid- March, however, cold tolerance decreased considerably as indicated by the smaller L150 values at both temperatures. These results seem consistent with Dickler's observations that cold tolerance remains approximately con- stant from October until February, but decreases by mid-March. In light of this fact and the smaller sample size of the test at -5° run on February 7, 1974, the December 17, 1974, results were presented in Table 81 and used in develOping the model. 117 .0000“ Locuo HHm :0 00 on we uemumc0 0mHHmmn ON 0Hm0050x00000 mo 0mH0200 co names mapmewpmm 00000000: « 0m.NH mm.0m m¢.om nm.om wN.00 mN.HN em.em mH.om mH.mm Ho.mm 0m.wN Nm.NN 0N.mH N0.m :o0uu0uw00 H000: 00.0H -- -- 0.00 -- 0.00 -- 0.00 -- -- 0.00 -- 0.00 -- 00. .0H .00: 00.0H 0.00H 0.00 -- 0.00 -- 0.00 0.00 H.00 0.00 -- 0.0H 0.0 0.HH 00. .H0 .000 00.H0 -- 0.00 -- -- 0.00 0.00 0.00 0.00 -- -- -- -- -- 00. .0H .000 H020000 00 00 00 00 00 00 0H 0H 0H 0H 0H 0 H 0000 oth 00° 00 meamoaxm 00:0: H0.0 00.00 00.00 00.00 00.00 00.H0 00.0H 0000000000 H000: 00.0 -- 0.00 H.00 0.H0 0.00 0.00 00. .HH .002 00.0 0.00 0.00 0.H0 0.00 0.0H 0.00 .00. .0 .000 00.0 -- 0.00 0.00 0.00 H.00 0.0H 00. .0H .000 00000000000 0 0 l, 0 0 0 H 0000 000. um mezmoaxu 00:0: .0000 0:0 00 0mEHH Hemem000u He metamoaxm 0000 2000 000000000 mHm>mH prHoueoz .00 mH000 118 Preconditioning_ A series of tests were conducted in February, 1975, to determine if a brief, non-lethal exposure to low temperatures caused any additional cold hardening beyond the level achieved from 6 months' storage at 40°F. In these tests, beetles in glass tubes were exposed in a low temperature bath for a period of time and then immediately transferred to a bath of an even lower temperature, and the mortality from the two exposures was determined. The time required for the beetles to change from one temperature to the other is considered a negligible part of the exposures. Table 85 shows the results of a test run on February 7, 1975, where 9 samples of 20 CLB's each were exposed to each treatment. In analyzing these results, two important factors are apparent. First, the constant exposures and con- trols showed mortality levels consistently higher than was predicted by the model, although only 1 of these differences was significant. This may, in part, be due to the seasonal tolerance change described in the last section. The other observation is that all 3 exposures with preconditioning re- sulted in less mortality than the unconditioned counterparts. The 2 exposures with 5 days preconditioning at 15° differed significantly (.01 level) from the unconditioned tests. However, the 6-hour exposure at 0° did not result in a significant decrease in mortality from the subsequent exposure at -5°. Since the mortality levels of the preconditioned CLB's in Table 85 include both the mortality due to the preconditioning exposure and the subsequent exposure, it must be concluded that the initial exposure was beneficial to the beet1es in significantly reducing total mortality from the 2 expsoures. Thus, preconditioning can be important in determining CLB mortality from cold exposure. 119 -- 02 00.0 00.0 00.0 -- 02 00.0H 00.0H 00.00 -- 02 00.0 00.0 00.HH -- .. 0H.00 00.0H 00.00 -- 02 00.00 00.0H 00.H0 e. -- -- 00.0H 00.00 .. -- -- 00.0 00.00 02 -- -- 00.H0 00.00 0000000000000 0> H0002 00 0000000000 .0.0 0H0H00002 00000000000000 0000: 0.0 :omm um>0000o 00>0H H0. 000 00 00000000000 .. mesmogxm oz 0O 00 00:0; 0 o0H H0 0000 0 .mHmummmm 00 pm 950: 2 O0. pm 0000: m 0000200000 000000000000: O0 00 00000 0H .00H 00 0000 0 00- 00 00000 0 .00H 00 0000 0 om- Hm 000°; m .oO 00 00:0; 0 0000500000 umcoHHHucoO .000200 000 0.0HO ON 00 0000200 0 mo cmpmwmcoo 00050000» .000000000000000 0005002 000 £003 mmgzmoaxm vHou 5000 0:000:00; mHm>oH zuHHmpeoz .00 00000 120 A second series of preconditioning tests was run on February ll, 1975, to determine whether the amount of preconditioning is influenced by the time and temperature of the preconditioning. Table 86 summarizes the' results of these tests, each of which consisted of 9 samples of 20 CLB's. An analysis of the results in Table 86 shows the following: l) 5 days preconditioning at 15°F was not significantly different from 6 hours at 25°F, but it was significantly different from 6 hours at 15°F. 2) 6 hours at 250 was significantly more effective in preconditioning CLB's than 6 hours at l5°. 3) 5 hours at 25° did significantly precondition the CLB's. Thus, significant preconditioning did occur in some treatments and the extent of preconditioning was determined by the time and temperature of precondi- tioning. Additional tests were run during the period of February 6 to February ll, 1975, to further quantify the relationship between time and temperature of preconditioning and the subsequent mortality resulting from a 3-hour exposure to -5°. Table B7 contains a summary of these additional tests as well as those already presented in Tables BS and 86. These re- sults, graphed in Fig. BS, show that there is apparently an upper level for preconditioning resulting in 20-25% mortality from the combined ex- posures. Additional preconditioning beyond the time needed to achieve this level resulted in greater mortality due to the increased lethality of the preconditioning treatment. Preconditioning occurs most rapidly at 25° where the maximum level is reached in 6 hours. At 15° it takes 5 days to reach about the same level. At 5°, preconditioning seems to occur at 121 Table 86. Mortality levels resulting from various preconditioning treatments. Each treatment consisted of 9 samples of 20 CLB's per sample. Means followed by the same letter are indistinguishable at the .01 level. Y% Mortality Treatment 6 hours at 15°, 3 hours at -5° 46.67 6 hours at 25°, 3 hours at -50 24.44 1 hour at +50, 3 hours at -50 47.22 5 days at 15°, 3 hours at -5°* 20.50 * Unconditioned, 3 hours at -5° 61.67 * From Table 5 16.39 13.10 17.87 8.54 16.58 122 Table B7. Mortality level resulting from various preconditioning treatments. Each sample consisted of 20 CLB's. PreconditioningTreatments Mor§a1ity S.D. Samples 1 hour at 5°, 3 hours at -5° 47.22 17 87 9 6 hours at 5°, 3 hours at -50 45.00 10.95 24 hours at 5°. 3 hours at -5° 81.67 15.27 3 1 hour at 15°, 3 hours at -5° 50.00 22.80 6 6 hours at 15°. 3 hours at -5° 46.67 16.39 9 72 hours at 15°, 3 hours at -5° 36.67 5.77 3 120 hours at 15°, 3 hours at —50 20.50 8.54 9 192 hours at 15°. 3 hours at -5° 33.77 14.07 3 6 hours at 25°, 3 hours at -5° 24.44 13.10 9 72 hours at 25°. 3 hours at -5° 23.33 5.77 3 120 hours at 25°. 3 hours at -5° 48.33 7.64 3 192 hours at 25°. 3 hours at -5° 53.33 20.21 3 Controls: 24 hours at 5° 21.67 5.77 3 120 hours at 15° 11.69 8.66 9 192 hours at 15° 21.67 7.64 3 72 hours at 25° 6.67 7.64 3 120 hours at 25° 23.33 11.55 3 192 hours at 25° 26 67 18.93 3 3 hours at -5° 61.67 16.58 9 No exposure 5.56 5.27 9 123 00-000000 .0020 .000000000000 000 00000 000000000 00 0000200000 000000000000000 000300000 00- 00 .000 0 we 00:0oax0 00 5000 000000000 000000002 .00 .000 owzocazoommm manor CON om. ow. 0.! ON. 00. on cm 000 ON o_w \\\U|IIIIIM.I\|\|I \\1o_ \\\ \ \ 00. m\\\ ... \ 60 .00 IIIIIIII III-IQ I ‘4 ‘ on. 4 .om.w H r. 500% 00?. ..V.. H l— .A o 40 ,"7 0 ID 4 .0002 3.20.00sz0... I I >CJ 0000 000>0P 000000002 00 000000000000000 00 0000000 000 .00 00000 126 mortality counts in all experiments were made 24 hours after the comple- tion of the exposure. The results of the comparisons of preconditioning and recovery are presented in Table 89, along with the mortality from an unconditioned group of beetles tested 5 days earlier. An analysis of these data reveals that in both cases the recovery treatments resulted in lower mortality levels than the 3 hours at -5° treatment, although only 1 of these differences is significant above the .05 level. A com- parison of the recovery treatments with the equivalent preconditioning treatments reveals that at 150 the beetles recovered significantly more than they preconditioned, and at 25° there was no significant difference between the 2 treatments. Recovery, unlike preconditioning, seemed to occur at the same level at 15° and 25° (31.11 vs. 32.22 % mortality). Thus, on the basis of this experiment. it seems that CLB's can recover from interrupted exposures; and, therefore, exposures are not necessarily additive. Furthermore, recovery and preconditioning are apparently different processes in the cerea1 leaf beetle. A series of experiments was conducted during January and early February of 1975 to quantify the aspect of recovery so that it might be included in the model. In these tests, CLB's were exposed to a repeated sequence of exposures and, periodically, samples of 60 beet1es were removed and mortality was counted. Most of the recovery periods were intended to cause minimal mortality, even when repeated several times. Thus, the mortalities shown in Tables 810 and 811 are caused primarily by the lower temperature exposures, and the recovery periods generally make an insignificant contribution to total mortality. The recovery periods are usually quite advantageous to the beetles, as even a lS-minute re- 127 Table 39. A comparison of preconditioning and recovery in reducing mortality caused by a standard exposure. Means followed by the same letter are indistinguishable at the .05 level. Treatment __X;__ _§;Q;_ 6 hrs at 15°. 3 hrs at -5° 46.67 16.39 a 1.5 hrs at -5°. 6 hrs at 15°, 1.5 hrs at -5° 31.11 8.58 b 6 hrs at 25°. 3 hrs at -5° 24.44 13.10 b 1.5 hrs at -s°. 6 hrs at 25°, 1.5 hrs at -5° 32.22 15.63 a,b 3 hrs at -5° 61.67 , 16.58 a 128 00.0H0 0.Hm 1- 0.00 0.00 1- ¢.H0 - -1 - cm 0; H .001 00; m N.o0HH 0H.000 0.00 11 0.00 H.00 11 0.00 11 11 11 00H :02 on .00- 0;; N N.mm0 N0.0Nm - 0.H0 0.00 0.00 11 0.00 - 0.00 - 00 :02 on .00- 0; H 00.H00 11 o.ooH 1- R.Hm 11 -1 1- 11 11 oo 00; wH .001 00; m Hm.0¢0 0.00 11 R.Hm 0.00 1- 0.00 11 N.me -1 oo 0; H .001 00; N 1- - -1 11 m.o~ o.o~ 0.00 0.HH 1- oo 0; H .001 0; H 00.000 0N.00~ 0.00 11 0.00 ~.m¢ 11 0.00 1- 0.0H 11 oo :05 on .001 L; H «0.000 00.0NH o.ooH 1- 0.0m 0.00 11 0.00 11 m.o~ 11 cc :05 0H .00- 0; H o.ooH 00.00 00.00 00.00 mm.om 00.00 00.00 00.Hm 00.0H 0000000000 00002 - 11 11 11 0.00 0.00 0.00 H.0N 0.0H mgamoaxm 0:000:00 agzogwmgzozom 00000010000000 0H m 0 0 0 0 m m H mucmzcmm mgzmoqu mama xgm>oumm xgm>oomm o0. 00 0000: .000 00- 00 mugsmoaxm 5000 xgw>ouwm .000 00000 129 H~.m0e 11 11 11 0.00 11 11 0.00 11 11 11 11 11 11 on «L; 0 .oo 00; 0H 00.000 11 11 11 o.ooa 11 0.00 11 11 0.00 11 11 11 11 00 00: NH .00 00; mm H.000 0.00 11 0.00 11 ".00 11 11 0.00 11 0.00 11 11 11 000 00; 0 .oo 00: oH ¢.¢¢N 00.000 11 11 o.H0 11 0.00 11 1- 0.00 11 0.00 11 0.00 11 omH 00: N .00 00; oH 0.000 00.H00 0.00 11 0.00 11 0.00 11 11 0.00 11 0.00 11 11 11 ooH 00; 0 .oo 0;: 0H 0.00 o¢.0oH 11 0.00 11 0.00 11 11 0.00 11 0.00 11 11 11 on 00; 0 .oo 00; oH 00.H00 11 11 11 11 0.00 11 0.00 11 0.00 11 0.00 11 11 oom 0;; 0H .00 mg: 0 00.000 11 0.00 11 0.00 0.00 11 11 11 0.00 11 0.00 11 0.00 00H 00: 0H .00 mg: 0 oo.ooH 00.00 00.00 00.00 00.00 00.00 00.00 00.00 00.H0 H0.~0 00.00 00.00 00.0H :o0uu00000 H000: 11 11 11 11 o.coH 1- 0.00 11 11 11 11 a.mH 11 mgamoaxu unnumcou Ngaocxmgaoxov Amusoz-00000ov 00 00 cm 00 00 00 mm on 00 ON 0H 0H 0 00:00000 ogzmoaxw 0000 000>ou0m 000>000¢ oo h< 00:0: .000 oo 00 0000000x0 5000 xu0>000z .HHm 00000 130 covery at 0° seems to increase survival from repeated 1-hour exposures at -5°. This is in spite of the fact that 0° is itself a lethal temperature to CLB's given longer exposures. The results in Tables 310 and 811 show that recovery is important, but require considerable analysis to determine to what extent and at what rates recovery occurs. For the purpose of including recovery in the model described in equations 1 and 2, the data in Tables 810 and 811 were ana- lyzed in light of this model. It is because of recovery that the state variable "exposure level" was included in the model. This variable is linearly related to time at any temperature and is deterministically re- lated to mortality by a probit regression equation. Thus, if after a sequence of exposures, a certain level of mortality is measured, the maxi- mum exposure level that the beetles were exposed to can be determined from the mortality level. In analyzing each experiment presented in Tables 310 and 811, exposure levels were calculated for each expressed mortality level. Then the addi- tional exposure level increase during the recovery period was calculated from equation 1 and subtracted from the "observed" exposure level. After developing a regression of this corrected exposure level vs. hours at the lower temperature, the slope of this line with recovery was subtracted from a similar slope without recovery, with the difference reflecting the amount of recovery (in degree-hours). Consider as an example the first treatment in Table 810. These beetles were given a 1-hour exposure at -5°, a lS-minute recovery period at 0°, then another -5° exposure, another recovery period, etc. It is apparent that this treatment resulted in greater survival than either the 131 experimental data or model predicted for constant exposure at -5°. To determine the extent of this recovery, the analysis shown in Table 312 was performed. The "observed" exposure level was calculated from the observed mortality using equation 2. The exposure level from the recovery period was calculated from equation 1. In determining the amount of re- covery, this additional exposure level resulting from the recovery period was subtracted from the "observed“ exposure level and a regression equation was calculated relating exposure level to hours at -5°. The slope (311.18 degree-hours/hour) was then subtracted from the corresponding slope of 431.94 degree-hours/hour which was calculated from equation 1 for constant exposure at -5°. The difference between these slopes (120.76) is the amount by which exposure level decreased during the recovery period. Similar calculations were made for recovery from each of the treatments in Tables 810 and 811, and these results are presented in both tables. When the calculated recoveries are plotted in Fig. 86 against recovery time, it appears that there is an upper level for recovery between 450 and 600 degree hours which is not exceeded by increasing the time or tem- perature of recovery. A comparison of the maximum recovery levels for -5° and 00 (considering all points between 450 and 600) reveals an insig- nificant difference between the means of 529.06 and 502.22 for -5° and 0°, respectively. Thus, the two temperatures can be grouped to give an average upper limit of 515.64 degree-hours for recovery. It is apparent from Fig. 86 that recovery rates are determined by the temperature at which, and from which, the beetles recover. With higher recovery tempera- tures, the beet1es recover more rapidly. They also recover more rapidly from -5° than from 0°. Additional data would be required to properly ana- 132 Table 812. Determining the amount of recovery during a repeated sequence of expsoures of 1 hour at —5° and 15 minutes at 0° (F). Observed Exposure Hours Observed Exposure Hours Level Observed Exposure Level at -5° Mortality Level at 0° from 00 - Exposure Level from 0 1 -- -- O 0 431.94 2 20.3 611.18 .25 16.91 594.27 4 53.3 1275.11 .75 50.72 1224.39 6 78.0 1776.00 1.25 84.54 1691.46 8 98.3 2775.33 1.75 118.36 2656.97 10 100.0 -- 2.25 152.17 -- y = 12.86 + 311.18x r2 = .98 431.94 - 311.18 = 120.76°Hours = Recovery 133 Am-¢¢mmm~ .mmzv ace 8:8 m- an megamogxm Faraway seem >go>oumg mo mgzpmgmasmu new «a.» mo co_uu==m a ma mpm>mp agm>ouwm o0 20mm Ew>oowm 950.... on- 20mm >mm>oowm manor m. N. w .v 0 00” I o O O— I q CON a O o 0 a D OWL a D 00 a O mwmnhdmmac‘m... >mw>oomm Av AV NV nu nu NV 0 m.— m. d 1 (>13 d d U u C) lC) It) 00. OON C) (3 ¢ (3* C) [O (SHOOH 338930) 13A3‘l AH3AOO38 CD C) 00 .mm .mrm l34 lyze recovery rates. however, for purposes of demonstration, this analysis will be continued. Since the maximum recovery level is reached in a relatively short time, it is only possible to calculate recovery rates from those treatments in Tables Bl0 and Bll which apparently did not exceed the time required to reach this level. 8y dividing recovery levels by recovery times, recovery rates (degree-hours/hour) were calculated and plotted in Fig. 87 vs. re- covery temperatures. These plots resulted in reasonably linear relation- ships for recovery from both -50 and 00 exposures. By sight-fitting lines to these points, a slope of 77.35 was calculated for -50 and l8.50 for 0°. These slooes reflect the rate at which the recovery rate increases with increasing temperature and, thus, have units of degree—hours/hour/degree. By plotting these slopes as in Fig. B7 and fitting an equation to them, it is possible to determine a continuous expression for recovery rates from any temperature at any temperature. In this case two points are in- sufficient for an adequate functional relationship so two likely alterna- tives are plotted. A mathematical expression for this relationship would give recovery rate as a function of recovery temperature and initial ex- posure temperature. It should be noted that although the data and statistical analyses show recovery to be a real biological phenomenon which exists independent of the model of eqs. l and 2, the quantification of this phenomenon is entirely dependent on the model. Discussion The sequence of developments in this modeling effort is apparent from the amount of data available on the various aspects. Initially, it was 135 Am-¢¢-mh .aazv n . .mgamonxm mszpmgmosm a c o o» mummguc. mgzpagmasou mo mmgmmu can >sm>oumg yo mums as» mcvpapmg covuucam Fasacmw w ”w” Wm>uw -mcgmppa mpnpmmoa N can mgsungngmu >gm>ouog mo covuucam a mo moo ucm m- sou» zgo>ouag mo mmuom .1. o mankdmwn—Zm... >mm>oomm O m... ON 9 0 ON . 0.6 . 32m 00 . >fi>8mm “a 83.0. «.4 cm ( 338930 / 8flOl-l / S8flOH o) 3d0'lS .0 20¢... wkdm rmmaomm mum. a.m.. 20mm mqu Ew>cowm 0.9 L o o N 8 O 0 d’ O O 0 000. CON. 0 O a) ( 338930 / 88001-1 o) 3.0118 A83A093 .mm .mwm 136 thought that a constant temperature model would be adequate when used with short time intervals in the field. In this light, considerable data was collected and the model of eqs. 1 and 2 was developed. It soon turned out, however, that exposures were not all additive and recovery had to be con- sidered. This aspect of recovery requires a great deal of data as three factors are involved: the temperature from which beetles recover, the duration of the recovery period, and the temperature of recovery. With additional data this aspect can be further quantified and a more comprehensive low tempera- ture exposure model can be developed, however it is difficult with the present data to develop even a preliminary model for recovery as it has not been discussed elsewhere in entomological literature, and there is no theory or experience to draw upon in modeling it. Preconditioning is an aspect which is more easily understood and modeled. It is apparent from Fig. BS that 6 hours of preconditioning at 25° is adequate to fully precondition CLB's and, thus, it seems reasonable to assume that in the field beetles are normally fully preconditioned when exposed to temperatures below 25°. In light of the preconditioning results of Fig. 85, it appears that some of the higher temperature ex- posures in Table Bl resulted in preconditioning of the beetles before any appreciable mortality occurred, but the lower constant temperature expo- sures killed them before they preconditioned. Thus, in its present form, the model is based on both preconditioned and unconditioned exposures. Additional tests should be conducted to develop time-mortality curves for fully preconditioned CLB's at temperatures from -50 to +5°F. These could then be used to recalculate eqs. l and 2 so they would predict mortality for preconditioned beetles. 137 The seasonal change in cold tolerance is an important factor which also requires additional data. Although it is now apparent that cold tolerance is reduced after February, this needs further quantification to determine to what extent tolerance is reduced and whether a further loss occurs as the season progresses. In the absence of these additional data, the model in its present form is limited in its utility. It is useful, however, in showing that even without preconditioning, CLB's can tolerate long exposures to moder- ately cold temperatures. Even at 5° it takes over three days exposure to reach 50% mortality. With the additional survival caused by precondi- tioning and recovery, it is apparent that CLB's are well able to tolerate the cold exposures they experience at the soil surface. Those beetles that overwinter above ground do receive low temperature exposures in the lethal range, and the current model might prove useful in determining upper limits for mortality in that its predictions are higher than occur in nature where preconditioning and recovery occur. APPENDIX C THE IMPACT OF RESISTANT WHEAT 0N POPULATIONS OF THE CEREAL LEAF BEETLE AND ITS PARASITES Introduction The cereal leaf beetle (CLB) has been observed by several authors to have preferences among its host crops of the family Gramineae. For oviposition, beetles generally prefer oats to wheat plants of the same age, and both crops are more suitable as host plants than native grasses. However, this preference is complicated by age of the plant, and if oats are as much as l0 days older than wheat, the preference is reversed (Wilson and Shade, l966). In Michigan, fields of oats and wheat, the primary host crops of the CLB, and acreages of native grasses, are frequently in close proximity and adult beetles distribute themselves among these three hosts. The synchrony of insect and plant development, which is determined by planting date and weather, is important in determining the distribution of the beetles. This synchrony, combined with the innate preference of the beetles, usually results in a relatively low CLB density in native grasses, a somewhat higher density in wheat (most of which is fall-seeded) and a much higher density in spring oats. An indication of the outcome of this interaction is the fact that between l969 and l97l, an average of 26.4% of the oats in Michigan were sprayed for CLB control, while only l.7% of the wheat was sprayed (Ruppel and Guyer, 1972; and Anonymous, 1974). 138 139 Shortly after the discovery of the CLB in North America at Galien, Michigan, in l962, a program was developed to breed crops resistant to the beetle. To date there has been little success in developing resistant oats; however, pubescent wheat has been found highly resistant to the CLB by in- hibiting oviposition, increasing developmental times, and increasing within- generation mortality. Schillinger and Gallun (T968) and Webster gt_al, (l973) compared the oviposition of CLB's on resistant and susceptible spring wheats in uncaged field studies and found reductions of 98.2% and 82.4% respectively, asso- ciated with the resistance. In a similar comparison with winter wheat, Webster gt_al, (l973) found a 96.3% reduction in oviposition on resistant wheat (CI 8519) compared with the susceptible Genesee variety. Of those eggs laid on the pubescent wheat, a high egg mortality was observed by Schillinger and Gallun (l968), and a high larval mortality was also observed by Schillinger and Gallun (T968) and Wellso (l973). Webster gt_al, (l973) did not observe a decrease in within-generation survival associated with pubescent wheat in the field. This observation was at variance with the other published work, and they attributed this discrepancy to probable contamination of their seed with a small amount of susceptible wheat, and to the decreased density of pubescence on the larger plants in their study. 1 0n the basis of these studies, it appeared that large-scale plantings of highly resistant wheat should greatly reduce or totally eliminate CLB production in wheat. This could have several important consequences on a CLB management program. If all wheat were effectively removed from the CLB environment, it is possible that those beetles which would normally 140 infest wheat, where they generally do not cause serious damage, would instead move into oats, thereby increasing the likelihood of spraying for their control. This could also work to the detriment of a biological control program since the wheat, which is seldom sprayed, serves as a sanctuary for CLB parasites. 0n the other hand, resistant wheat could work to the advantage of a biological control program if it encouraged additional oviposition in wild grasses where the larval parasites would be less subject to the deleterious effects of plowing while diapausing in the ground in CLB pupal cases. The object of this study, conducted in l972 and l973, was to determine the impact of large plots of resistant wheat on the cereal leaf beetle and its larval parasites. Since it was felt that the high level of resistance provided by the pubescent wheat was possibly more than was needed in a CLB management program, plantings of mixtures of resistant and susceptible wheat were also evaluated. Methods Plot Description On October l6, 1972, a field at the Michigan State University Kellogg Biological Station Research Farm in Kalamazoo County, Michigan, was planted with resistant (R) and susceptible (S) wheats and mixtures of R and S as shown in Fig. Cl. The resistant germplasm called Vel (alluding to its velvet-like pubescent surface), was developed and propagated at Purdue University CI l5890 where it was selected for pubescence. The susceptible wheat, Genesee (CI l2653), is a soft white winter wheat commonly grown in Michigan. All the plots were seeded at a rate of 2 bu/acre and fertilized with 250 lbs/acre of 6-24-24 at planting. The combinations of resistant 141 Asp-e¢mums .mmzv .ummm m new a 0o mmgsz*s mcwcpmucou mpopa mg» new m . . uoPa m x m pamumzm an ax noopv acmpm*mmg ms» mo covumuop mg» m:_3o;m umvuapm muopa poms: ummummwmgpmmu mo “30mm; llqlrt‘. om mflo .: con m 800. mhdo talc. zoo. zoo. .: oon .: cm. 5.8 5.8 3.9 58 589 _ m._.osmg .EEV mmmEcu mcwvmmm stn< .mu mpnmh 149 The densities of adult CLB's in the 100% R and 0% R plots as deter- mined from the sweepnet survey are presented in Table C4. The densities in the l00% R and 0% R plots are plotted in Fig. C2 along with the aver- age egg density in the 5 plots. This graph indicates that the densities in these plots of almost pure susceptible or resistant wheat did not differ appreciably until after the onset of oviposition. Samples taken at peak adult density on May l8 indicated significantly more CLB adults in the 0% R than l00% R plots (.05 level with a t-test). Subsequent samples indicated no significant differences between the plots as densities diminished with time. The average adult CLB densities in the l2 oat fields on the Gull Lake Farm are included in Fig. C2. This graph of oat densities indicates that at the time the differential adult density in the l00% R and 0% R plots developed, CLB's were beginning to move into the new oats, their preferred host. Thus, the beetles which would normally have gone into the R wheat could have instead gone into nearby oats. The larval densities measured with the sweepnet survey are presented in Table C5. In Fig. C3, these densities are plotted against degree-days > 48°F. The area under each of these curves was divided by 2400 days, the larval developmental time used by Tummala gt a1.(l975), to determine the number of individuals produced per unit area for the season (Southwood, l966). From these numbers included in Table C5, the % reduction from the 0% R plot was calculated for each plot. The "pure" R wheat caused a reduc- tion of 83.2% in the number of larvae produced and the mixed plots (90% R, 50% R, and l0% R) caused reductions of 74.4%, 47.5%, and l.3%, respectively. Wellso (l973) observed an increased larval developmental time of about l0% 0n seedlings of a pubescent wheat (CI 85l9) compared to the S variety, 150 Table C4. Adult cereal leaf beetle densities per 1000 square feet as determined by a sweepnet survey. PLOTS Date 100% R 90% R 50% R 10% R 0% R April 24 8.3 8.3 12.4 0 1.7 May 1 7.6 7.6 7.6 30.6 7.6 May 7 40.0 32.0 48.0 40.0 8.0 May 9 72.5 46.4 98.6 116.0 64.0 May 18 86.0 16.8 63.0 67.2 210.0 May 24 47.5 34.8 92.8 63.8 98.6 June 1 27.5 36.0 55.4 0 34.3 June 7 14.0 0 10.4 23.4 4.7 June 13 0 0 0 0 0 151 Table C5. Cereal leaf beetle larval densities per l00 square feet as determined by a sweepnet survey. PLOTS Date °D>48°F 100% R 90% R 50% R 10% R 0% R June 1 5l0 l l 0 0 0 June 7 64l 27 37 78 105 140 June l3 796 l5 33 62 149 116 June 20 968 3 3 0 4 9 June 27 1119 0 0 1 0 1 July 5 1292 o 0 o o 0 # Produced/loo ft2 28.3 43.1 88.3 166.1 168.3 % Reduction from 0% R 83.2 74.4 47.5 l.3 0 152 A¢-e¢-m~ .mmz .cOmamm ecu usogmzogga magma: mpavpnmumzm ecu “coumwmwg mo mpo_a umxms can mean 65p a? mmwpmmcmu P~>gm4 .6. .90 x m> 48, the % reduction in the 100% R increases to 84.7%. Cereal Leaf Beetle Behavior The results of the observations on beetle movement (Table C6) in- dicate behavioral differences between the beetles in the 2 wheats. On May 20, the beetles were found to fly farther in the susceptible than in the resistant wheat. This difference was found to be significant (P < .99) with a t-test. None of the other behavioral differences between beetles in the two plots in Table C6 proved to be significant, but the beetles consistently took shorter, more frequent flights in the l00% R plot than in the 0% R plot. From these data on flight distances and frequencies, a diffusion coefficient can be calculated as: 2 D = 1%%%—- (Pielou l97l) where t = displacement in a plane (= flight distance). The diffusion co- efficients in Table C6 describe the rate of spread of a p0pulation across a field as a result of random movement. Clearly, the more rapidly a popu- lation distribution is spreading out, the more rapidly individuals are getting out of a field of finite size (assuming non-reflecting boundaries). Since on May 20 the beetles had a higher diffusion rate in the resistant wheat and 6 days later it was higher in the susceptible, it appears that additional work will be required before generalizations can be made about the impact of resistant wheat on diffusion rates. 0n the basis of Fig. CZ it appears that this might be a dynamic relationship and differential dif- fusion might only be apparent during oviposition. 154 mm.omp oF mm. NP om.n m ROOF 3N.mm NF .m.=mo._ m .m.=mn.oF a fig om >62 cm.on 58 Pm. Fm mm.m m ROOF mm.omm mm .m.cmm. mm RRFN.PP a &0 ON has A.:Ps\ swv mcowpm>gmmno A.cmsv mcowpo>gwmno ~.:wv pofia mpmo “cave? emou memh.m mucmumwo.m coymsmmvo .pmmcz acmpmwmmg new mpnmpgmomzm cm mmpummn wme megmu com mpcmwp+ :mmzpmn mpm>gmpcw vcm mmucmpmwu unmmpm co mcowum>gmmao .mu man» 155 Caged Studies By caging ovipositing adults in the 6.6 ft. square plots for l week, a pulse of eggs was introduced into the plots. While it is not very meaningful to compare the egg production of confined adults in different plots, certain aspects of egg distribution and within generation survival after the cages were removed are of interest. Distribution within cages. On May 29, egg counts indicated a con- siderable variation in density within individual caged areas, but an aver- age distribution that was almost equally divided among the 4 sq. ft. subplots in the corners of the milliacre caged areas (26%, 23%, 27%, and 24% in the NW, NE, SW, and SE corners, referred to as subplots l-4 in Table C7). The non-preference for R wheat for oviposition, which was ob- served throughout the field (Table C2) was also indicated in the caged plots in mixed R and S wheat where 97%, 98%, and 92% of the eggs were laid on the susceptible plants in the l0% R, 50% R, and 90% R plots, respectively. This indicates that the beetles sought out the susceptible plants in the plots for oviposition and as a result deposited an average of 95.7% of the eggs on the S plants in the 3 plots. The larval count of June l2, Table C7, indicates a shift in the dis- tribution of beetles in the caged plots from the time of the egg count because an average of 84.7% of the larvae were found on the S wheat (100%, 79%, and 75% on the S wheat in the l0% R, 50% R, and 90% R plots, respectively), compared to 95.7% of the eggs. This redistribution could be due to higher mortality on the S plants or movement from S to R plants. Withinegeneration survival. The number of adult CLB's produced in each caged subplot as determined from the emergence trap catch is presented 156 in Table C7. These adult densities were divided by the egg densities in the same subplots to determine the within-generation survival. Helgesen and Haynes (1972) showed that CLB within-generation survival is inversely related to the log of egg density, so it is not possible to directly com- pare survival in plots with differing initial egg densities. Since in this test the 3 plots with mixtures of R and 5 plants had similar average densities of eggs [44.0, 49.5, and 43.5 in 10% R, 50% R, and 90% R, re- spectively (Table C7)], the within-generation survivals in these plots are directly comparable. The average survivals of .1059, .0896, and .1408 in these plots are not significantly different and show no consistent relationship between within-generation survival and relative percentage of S and R plants in the plots. As a result of large differences in egg densities in the 0% R and 100% R caged plots (X'= 53.3 and 24.9 per ft., respectively), the survival in these plots is compared by covariance analysis. In Fig. C4, a plot of survival vs. the log of initial egg density indicates that in both 100% R and 0% R plots survival decreased with increasing egg densities. The slopes of both lines differ from 0 (P > .99 in S and P > .90 in R). The slopes of the 100% R and 0% R lines are not significantly different from each other; however, the means of the covariate (egg densities) are signif- icantly different (P > .98). Thus, % survival was adjusted for the effect of egg density. The adjusted means (8.33 and 12.98% survival in the 100% R and 0% R wheat plots, respectively) are not significantly different at the 5% level. In summary, there was density dependent mortality in the plots with "pure" R and "pure" S wheat; however, when corrected for dif- fering egg densities, within-generation survival in the two wheats was 157 Table C7. Cereal leaf beetle egg and larval densities and total emergence trap catch in each of the 4 square-foot subplots (densities are per square-foot). CLB's CLB Plot Introduced Subplot Eggs Larvae Adults 1, julis Adults/Egg 0% R 800 1 48 39 5.50 3.00 .1146 2 28 19 4.25 3.50 .1518 3 124 80 7.00 1.50 .0565 4 44 56 5.50 1.00 .1250 400 1 67 37 3.25 0.0 .0485 2 44 23 6.25 3.00 .1420 3 48 31 4.00 1.75 .0833 4 23 18 3.50 1.75 .1522 10% R 400 1 50 37 6.25 1.25 .1250 2 39 11 4.75 .25 .1281 3 39 26 4.00 1.25 .1026 4 48 19 3.25 .25 .0677 50% R 400 1 33 15 3.75 1.00 .1136 2 67 29 3.25 1.50 .0485 3 26 34 3.75 .50 .1442 4 72 45 3.75 2.00 .0521 90% R 400 1 42 10 4.50 0.0 .1071 2 62 18 4.25 .75 .0685 3 30 11 5.25 .25 .1750 4 40 13 8.50 .25 .2125 100% R 400 1 7 7 1.25 0.0 .1786 2 8 10 .50 1.00 .0625 3 7 B 1.00 1.25 .1429 4 B 24 2.25 1.75 .2813 800 1 62 17 1.50 .50 .0242 2 19 6 1.25 .75 .0658 3 39 10 2.50 .50 .0641 4 49 13 4.50 2.00 .0918 158 M5133~Nm~ .mmzv .zpvmcme mam Passes. co corpoczw a ma paws; mpnwuamumsm mgaa can acmpm?mog mg: :3 a.mgu mo ~m>w>gam cowumgmcmm cwguF2 .vu .mwu >._._mzwo 00m OON oo. 00 0* ON 0. o v N - 4 - J a q a q 4 q 1.... 7.. nxv nu .40. 8 .An m L m. HI. . .1. 3. an. 18 0 :Nn + vac—0.0. 1. or V a we BIG 0 N¢.uNL lmN .m omNN + xvnNd... vi 1. a «con: mwuuAu nu zoo ...un 159 similar. Since these survival rates are measured in populations of unnatural age distribution (i.e., resulting from a pulse of eggs), their utility is limited to comparisons and not pr jections for natural popula- tions. Tetrastichus julis. In Table CB, it can be seen that there are large differences in the number of pupal cells in the 100% R and 0% R plots (Y'= 13.38 and 27.75, respectively). However, there is no apparent relationship between density and percent parasitism at this range of densities, so it is possible to directly compare the % parasitism in the different plots. This comparison reveals an average parasitism of 49, 41, 48, 49, and 46% in the 100% R, 90% R, 50% R, 10% R, and 0% R plots, respectively. None of these differences are significant, so apparently the attack rate of I, juljs is not affected by either pure resistant wheat or mixtures of R and S. I, julj§_is a gregarious parasite with a facultative diapause (in contrast with L, curtus_and Diaparsis n. sp. which are solitary with a facultative and an obligatory diapause, respectively). A comparison of the percent diapause of of I, julj§_reveals no significant difference between the 100% R and 0% R plots. Also, there is no difference in the number of parasite larvae per pupal cell in the 2 plots. Thus, it ap- pears that R wheat has no direct effect on T. julis. Othergparasites. The densities of Lemophagus curtus and Diaparsis n. sp. were very low in 1973 and as a result there was a low parasitism rate by these species, as shown in Table C8. These small numbers preclude any significant analysis; however, they give no indication of an effect of R wheat on either parasite. There was an average parasitism by L, curtus 160 Table C8. Evaluation of cereal leaf beetle pupal cells in the 4-ft2 subplots. .am .c mwmgommwo cur: mppmu mngsu 24 £823 appau aw 2H :82: mfi_mo Peace mm>LmN aw aH mcwmamamwo aw aH mewmzmamwo ;p_3 mfipmo .n aH nmmLmEm cuwz mFqu m_m40 ummo m.m4o vmmgmsm m_Pmu Payee uo_afl=m umuzuocch m.mgo uo—a 0000 1.120 0001.. 0210 12 14 6 11 29 31 14 30 32 37 4433 00.10 0000 14 16 24 8 24 9 10 11 5 4 10% R 400 0000 1010 4975 2141 11.00 1234 50% R 400 0000 3022 5234 90% R 400 0100 0002 4731 18 4531 0200 0011.. 4834 2000 001.1 7 6 14 10 9 15 26 27 3468 4282 1020 8475 161 of 3.7% in the 100% R plot vs. 3.2% in the 0% R plot which is remarkably close considering the small sample size. Diaparsis was found to parasitize larvae in both plots--however, in numbers too small to compare. Discussion The results of this study on the impact of R wheat can be summarized as follows: 1. CLB's oviposit less on R wheat than S, in both pure stands and in mixtures. 2. CLB's exhibit density dependence in within-generation survival in both R and S wheat. 3. When this density dependence is taken into account, there is no difference in within-generation survival on pure R or S, or mix- tures of R and S wheat. 4. The parasite I, jglj§_is not directly affected by R wheat, although by reducing CLB oviposition, the R wheat seems to reduce the number of potential parasite hosts in wheat. Thus, the primary effect of R wheat is to reduce oviposition; however, since densities are lower, within-generation survival can be higher in R than S wheat. Because the relationship between density and within- generation survival is not affected by R wheat, the model of Helgesen and Haynes (1973) can be used to predict within-generation survival in S, R, and mixtures of wheat. Also, since R wheat causes no direct effects on I, juljs, the model of Tummala et_a1, (1975) can be applied to R, S, and mixtures of wheat. 162 The R wheat apparently retained as many adult CLB's as S wheat until oviposition began (Fig. CZ). It appears that the beetles which left the R wheat during peak oviposition did not move to native grasses or sus- ceptible wheat, but to nearby oats, the preferred host, which was up at that time. However, as seen in C2, the R wheat maintained a significant CLB population even during oviposition, so in a future management program, the impact of beetles moving to oats from R wheat may not be too severe. Also, a slight adjustment of the planting date could result in oats not yet being out of the ground at the time of this movement and more beetles might move to native grasses. A large-scale release and utilization of R wheat would apparently result in a large reduction in oviposition in wheat. As a result, CLB damage in R wheat would not be significant, and numbers of CLB's produced in the wheat would be greatly reduced. During the first season, some of those beetles which would normally oviposit in S wheat might be expected to move to nearby oat fields. However, in subsequent years, the small number of CLB's produced in R wheat should not significantly affect the oat populations. Under current planting practices the three larval parasites would be adversely affected by a large scale release of R wheat. The ovipositing beetles, denied access to the wheat, might be expected to move to native grasses, however if oats are available, as they were in this study, the beet1es will most readily move to the young and highly preferred oats. As a result, a greater proportion of a regional population of CLB larvae and parasites would be found in oats where they would be subject to in- secticide application. Although the temporal and spatial availability of 163 oats is an important factor in a particular year, it appears that in general the loss of the CLB and parasite refuge in wheat would result in a greater degree of instability in the system and an increased reliance on chemical control. The resistant wheat variety Vel probably has a higher level of resistance than would be required in most areas in a CLB management pro- gram. For a regional program, several R wheats with a range of resistance levels would be desirable, or the resistant seed can be mixed with its susceptible counterpart to achieve the optimum level of resistance. APPENDIX D THE STRIP SPRAY COMPUTER SIMULATION USED IN APPENDIX A 164 165 Program Strip [INITIALIZE PARKIERS ] SUBIOUTIHI DIFFUSE CALL DIFFUSE (Cale. : loving each 0 of feet) y i J SUBIOUTINE HURT CALL MONT .17 (Cale. I mortality from each El) PER I 00 4 I-l. N(N-Minutes Simulation) ] "L . SUBIOUTINE LOCATE CALL LOCATE Calc. I of each E1 in each foot of strip) I SUIRWT INE EXPOSE (Update El of CLB': on insecticide) J CALL EXPOSE 2 EACH EL 2 EACH FOOT I J l 4 CONTINUE 166 PFCUURQN $3TR'IF' (:[NF'UT‘3128 7 01111311153313.2113) 1:121:1“11311‘1533201‘1 X ( 1340) 7 PER ( 200) 9 Z( 13001440 ) 7 Y (19.00.. 40) 71.-01.21.1030 ) 1: ‘1” LUTC’S 1' 1,3300 7 14118335000 L:;‘:1k>§i>k3¥3¥>f1<>1<>’.<>1i< U 13" III {-1 T138 E X F' (.1 531.1 111' E 1... E: U E 1.- ( 13.1.. ) {'11: EXxEx~5. AXmABS(EX) TT=1./(1.+.2316419*AX) nmo.398942385XP<~EX*EX/2.1 $3111.”. I 151 '1' 1331.1 T (.1 F' E 151‘ C 13. N '1' 8 1x112 TH EACH EL F'=Il. . ~[I$11<(( ((1 .330274X'TT-r1 .821256)*1"T+1.781478) +)11<'1'T+0 . 3193815) IF(EX)39292 P=1e*P 13'1331'\"(1' 1:1:'*21()()o CONTINUE RE T U RN ENIJ $331.113‘R'OUTIT‘1E LOCATE ( NT? 7 NT 9 X r Y 9 Z) IIIMIEINSZL'ON 2(200940) 9Y(200940) rX(",-.’40) DO 1 L=17200 DU 4 I=11NT Z(LyI)zY(LyI)$X(1) MmI+1 IF(M.GT.NT)M¢1 DO 5 JWRPNTR Z(L9I)=Z(L!I)+X(J)*Y(L7M) M3M+1 IF(M.GT.NT)M31 CONTINUE Mwlwl IF(M.LT.1)M&NT [IO 6 J==112 7 NT 2 Z(LvI)=Z(L!I)+X(J)*Y(L9M) HxMul IF(M.LT.1) MmNT CONTINUE CONTINUE CUHTLHUE RETURN END SUBROUTINE EXPOSELNTyNUvELvaZ) DIMENSION 2(200140)7Y(200940) N.1.'»=‘-NU+.1. NL==|§§L+ . 1'7] [IO 1 I-“éIyRC/O I 1‘ iii 155: C '1' 53 (.1 1‘1 IN EACH FT 4 .. .. .. OF THE STRI F’ I 10.813313 '1' I C I [11:3 J) 169 U0 4 K=1rNU Y(IIK):Z(I!K) CONTINUE KNmI+KL HO 2 J=N1vNT IF(KK.LT.200) GO TO 3 Y(QOO!J)3Y(2009J)+Z(19J) GO TO 1?. Y(KKvJ)=Z(IyJ) CONTINUE CONTINUE DO 5 1:1!200 DO 6 J=19NT Z(IvJ)=Y(IvJ) CONTINUE CONTINUE RETURN ENU APPENDIX E STUDY AREA SURROUNDING THE KELLOGG GULL LAKE RESEARCH FARM, KALAMAZOO COUNTY. ROSS TOWNSHIP 17D 171 SECTION 4 r--—-----------—-----------------—- ------------1 1 -—----—--J F1 1 9 B _ - m I - 2 _ 4 I I 1' .. TIIIJT B. 2 h n Tu . 3 6 _ 2 2 — Lu On —n «n n . m 1.111111L1._1-J- 172 SECTION 5 r-----------------------------------—-----------‘ J‘U'U' ' ""'I"--" -"-'-""'I' 'I"""'."‘ B u . O . l - - - II 2 3 - I. .l I - 9 . — - 5 . I - - c 7 - T 3 J l O a. n A 3 3 N O. I 6 On b J .n b 2 J. on . 4 v« —m" “u 3 u 3 u 3 an I : MB 11.. a M... 4 . .2. I B “u 3 - o: 3 1“ o. u - on J u 7 6 n 2 0 3 3 a . C. J a." 9+ O + I n 3 4 4 4 4 5 u n 4 4 - 8 u on 3 . 9 2 n 4 4 . I. . 8 7 - 4 T 4 . LU-"|""'I""I' 'I"'r"-"ll'l""""' '- 173 SECTION 8 .n I .1 _ 7 I B .I in T i T T 74 ) ’4 O w ||'II|IIIL 174 SECTION 9 ----------1 I 4 5 T SCHOOL T |—-—_