tga'a‘gziéeizzrykzr‘nec-.r=:_I ,1. I- : . A ‘ ‘ \fi‘lkflqhflpl .i...~ ”1:. . THE INTEGRATION OF ADULT SURVIVAL AND DISPERSAL INTO A MATHEMATICAL MODEL FOR THE ABUNDANCE OF THE CEREAL LEAF BEETLE. Oulema melanopus (L) Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY. WILLIAM G. RUESINK 1972 MM LIBRARY lflhflfigznfihnu: University This is to certify that the thesis entitled THE INTEGRATION OF ADULT SURVIVAL AND DISEEESAL INTO A MATHEMATICAL MODEL EOR'THE AEUNDANCE OF THE CEREAL LEAF BEETLE, Oulema melanopus (L.) presented by William G. Ruesink has beentaccepted towards fulfillment of the requirements for Date 4/7”, 3’“ 0-7639 - ‘J * amome av :_ HDAB & SUNS' I fiflflflflflfi ABSTRACT THE INTEGRATION OF ADULT SURVIVAL AND DISPERSAL INTO A MATHEMATICAL MODEL FOR THE ABUNDANCE OF THE CEREAL LEAF BEETLE, Oulema melanopus (L.) By William G. Ruesink The main objective was to develop and analyse a systems model for the population dynamics of the cereal leaf beetle, Oulema melanopus (L.). Numerical values for the parameters were, in general, taken from the literature; however, nothing was available on the mortality rate for the adult beetle nor for the time varying spatial distribution of adults among the various habitats. Consequently, field work was designed and executed to investigate these features. The resulting model has as components the life stages of the insect: egg, 4 larval instars, pupa, and 3 somewhat arbitrary sub- divisions of the adult stage (summer, overwintering, and spring). The internal structure of each component is essentially an accounting of the number of individuals moving in and out together with a time lag ("developmental time") that is a function of environmental temperature. Analysis of the model revealed its response to a variety of stimuli, one of the most interesting being fluctuating temperatures. For example, rapid buildup of the beetle pOpulation is favored by cool springs, especially when there is a large difference between day and William G. Ruesink night temperatures. Hence the predicted buildup rate for Alpena, Michigan is about twice as great as for Lexington, Kentucky. A second type of response is caused by the differences in temperature between days. Acting through the oviposition rate and developmental times, this fluctuation causes large day to day differences in larval density; changes of up to 20% in a single day are common. The mortality rate for adult beetles was studied in cages and by a regional survey. Although the magnitude of the mortality rate remains in doubt, it appears that the average daily temperature has a major effect on that rate. For the purposes of the model spring adult mor- tality was taken as 0.2% per degree-day (above the base of 48° F.) prior to May 18 and 0.4% thereafter. Summer adult mortality was taken as 0.05% per degree-day during the feeding period. Overwintering mor- tality was taken as 50% between the time summer feeding ceases and spring emergence occurs. Movement of spring adults between habitats was studied using traps to catch the emerging adults followed by a survey of grain fields. Emergence from overwintering sites occurs primarily from 50 to 150 degree—days, which may take up to 25 calendar days. A portion of the emerging population soon moves into winter grains; the remainder appar- ently move around the environment and await the occurrence of spring grains. Furthermore, it appears that those beetles entering wheat stay there rather than later moving to oats as has been reported in the literature. Movement of summer adults into overwintering sites was measured primarily by sifting beetles from soil samples using a cotton gin trash William G. Ruesink mill. Samples taken from August through November indicate that progres- sively fewer beetles can be found in the top 3 inches of the soil. This, combined with the fact that in August the measured population was comparable to the population measured the following spring, implies that the majority of the beetles probably overwinter deeper than 3 inches in the soil. THE INTEGRATION OF ADULT SURVIVAL AND DISPERSAL INTO A MATHEMATICAL MODEL FOR THE ABUNDANCE OF THE CEREAL LEAF BEETLE, Oulema melan0pus (L.) By xx. William G? Ruesink A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Entomology 1972 ACKNOWLEDGEMENTS Perhaps every student upon graduating feels much as I do about his major professor. Its only natural, I presume, since he is the one person whose professional interests are closest to the student's. Also there is no other faculty member with whom the student has had greater contact. Dr. Dean L. Haynes has been more than an advisor to me; he has become the standard of quality against which I will compare all other entomologists and their professional efforts. There is no single trait about him that makes me feel this way, instead it is the combination of his insight, imagination, dedication, foresight, preseverence, and humanness. No other single person comes close to having as much in— fluence of my attitude toward the profession of Entomology. Although my debt to Dr. Haynes is by far the greatest, I also owe special thanks to: Drs. Gordon Guyer, Kenneth Cummins, William Cooper, and Herman Koenig for serving on my guidance committee; Drs. Robert Ruppel, James Webster, Stanley Wellso, and Mr. Richard Connin for sharing information on the biology of the cereal leaf beetle; and Dr. Gerald Park for teaching me the elements of, and whetting my in- terest in, systems science. Finally, I wish to acknowledge the help of everyone in the Department of Entomology, because I honestly believe that half of what ii I know about practical entomology came not from classes or thesis re- search, but from personal conversations with individuals in informal situations. iii TABLE OF CONTENTS INTRODUCTION . . . . . . . . . . . . . . . LITERATURE REVIEW . . . . . . . . . . . . DESCRIPTION OF THE MODEL . . . . . . . . . ANALYSIS OF THE MODEL . . . . . . . . . Refining the model Sensitivity of the model Predictions and interpretations Initial density Portion preferring wheat Available host acreages Temperature distribution Daily temperature fluctuations bIETHODS O O O O O O O O O 0 O O O O O O O The study area Overwintering sites Cage studies of adult mortality Population survey RESULTS . . . . . . . . . . . . . . . Overwintering sites Rate of emergence from overwintering Cage studies of adult mortality Regional population survey DISCUSSION . . . . . . . . . . . . . . . . Overwintering sites Adult mortality from survey data Adult migration from wheat to oats CONCLUSIONS . . . . . . . . . . . . . . . LITERATURE CITED . . . . . . . . . . . APPENDIX 0 O O O O O O O O O O O O O 0 iv 35 39 59 66 68 71 Table 10. ll. 12. LIST OF TABLES Mortality Rates for the Immature Stages of the Cereal Leaf Beetle O O O O O O O O O O O O O O O O O O 0 Comparison of MOrtalities Measured by Several Authors at a Common Density of About 200 Eggs per Square Foot in Oats . . . . . . . . . . . . . . . Days Required for Each Instar to Complete Larval Development at 60°, 70°, and 80° F. . . . . . . Effect of a .10 Error in Any Particular Survival Esti- mate on the Resulting Deviation of Generation Index (I) from Its Accepted Value of 1.08 . Average Daily Temperatures for April Through August at Selected Locations in the Potential Range of the Cereal Leaf Beetle Generation Index (I), Fecundity (F), Summer Adult Survival (S), and the Dates for Spring Adult Emer— gence (SP), Peak Larval Density (PL), and Summer Adult Emergence (SU) as Predicted for 5 Locations . August 11-18, 1969 Gin Mill Finds by Habitat at Gull Lake . . . . . . . . . . . . . . . . . Late September and Early October 1969 Gin Mill Finds by Habitat at Gull Lake . . . . . . . . . . . . . . October 23 - November 6, 1969 Gin Mill Finds by Habitat at Gull Lake . . . . . . . . . . . . Overwintering Mortality in Above Ground Habitats; Gull Lake 1970 - 1971 . . . . . . . . . Catch in Emergence Traps for Overwintering Adults 1971 O O O O O O O O O O O l O O O O O O O O 1971 Spring Adult Emergence by Habitat at Gull Lake . Page 21 30 31 4O 41 42 44 45 47 Table Page 13. 1971 Spring Adult Emergence by Date at Gull Lake . . . . . 49 14. Adults Mortality as Computed from the 1970 Cage Study . . . S4 15. Number of Cereal Leaf Beetles in Grain Fields in the 4 Square Mile Study ARea for 1970 . . . . . . . . . . . . 55 16. Number of Adult Cereal Leaf Beetles in Grain Fields in the 4 Square Mile Study Area for 1971 . . . . . . . . 56 17. Spring Adult Mbrtality as Computed from the 1971 Regional Population Survey . . . . . . . . . . . . . . . 63 vi LIST OF FIGURES Figure Page 1. Components of the Cereal Leaf Beetle Population Dynamics Mbdel . . . . . . . . . . . . . . . . . . . . . ll 2. Generalized Internal Structure of a) the Spring Adult Component, and b) All Other Life Stage Components . . . . 12 3. Predicted Influence on the Generation Index of Varying a) the Initial Population Density (D) of Spring Adults, b) the Portion (P) of the Population Pre- ferring Oats Over Wheat, and c) the Portion (R) of the Small Grain Acreage that is Planted to Oats . . . . . 27 4. Two Examples of Predicted Day to Day Fluctuations in Larval Densities: a) for Lansing, and b) for Alpena, Michigan . . . . . . . . . . . . . . . . . . . . 34 5. Cumulative Emergence from Overwintering Sites at Gull Lake for 1971: a) Over Calendar Date, and b) Over Degree-Days Accumulated Above the Base of 48° F. . . . . . . . . . . . . . . . . . . . . . . . . 50 6. Log-Probability Plot of the 1971 Gull Lake Cumulative Emergence from Overwintering Sites . . . . . . . . . . . 51 7. Rate of Emergence from Overwintering Sites at Gull Lake for 1971, Computed from the Observed Cumulative Emergence O C O O O O O O O O O O O O O O O O O O O O O O 52 8. Pupal Survival in Oats and in Wheat at Gull Lake for 1970 O O O O O O O O I O O O O O O O 0 O I I O O O O O O 58 9. Estimated Spring Adult P0pulation in the 4 Square Mile Study Area for 1971 . . . . . . . . . . . . . . . . 62 vii INTRODUCTION Population dynamics is the study of birth, death, migration, and developmental rates. Considerable work has been done on the population dynamics of the cereal leaf beetle, particularly by entomologists at Michigan State University, Purdue University, the Commonwealth Institue for Biological Control in Delemont, Switzerland, and the Canada Depart- ment of Agriculture in Harrow, Ontario; all of the pieces fit together to form the total story. One of the best methods available for tying together many diverse observations is the method of systems modelling and analysis. Within the field of systems science there exist two major approaches. The one uses models that are mathematically tractable and manipulates these models using standard mathematical procedures. Predictions are obtained from the model in concise form: that is, the answer to any question posed comes out as a simple factual statement. The other approach is to build a logically sound but mathematically complex model that can only be analysed by using large computers. Even then results normally require graphical presentation to be comprehensible. Each approach has its advantages, but essentially in the first the modelling is difficult and the analysis simple, while in the second the Opposite is true. In early 1971 entomologists at Michigan State University joined with systems science engineers as part of a National Science Foundation project (GI—20) entitled "The Design and Management of Environmental l 2 Systems". Together they are modelling the life system of the cereal leaf beetle with the goal of designing an effective pest management program to minimize the effect of the beetle on small grain production. The model presented in this thesis was essentially complete before this cooperative effort began, and consequently it served as a prototype for their efforts. In the process of developing this model of the population dynamics of the cereal leaf beetle, it became clear that certain aspects of the beetle's dynamics were not adequately described and quantified. Field work was undertaken to do the following: 1) quantify the adult mor- tality rate as it varies throughout the year, 2) describe and quantify the movement of adult beetles between habitats, and 3) quantify the relative use of various habitats as overwintering sites. The results of the field work, although not as definite as one might hope, were adequate to complete the development of the systems model. LITERATURE REVIEW Details of the cereal leaf beetle population dynamics are needed for the development of the systems model in the next section; hence all relevant data from the literature are given here. Oviposition rate has been studied in the field (Helgesen, 1969) and in the laboratory (Yun, 1967). Maximum daily temperature (T) cor- relates highly with eggs per female per day (E) over the temperature range of 55° to 80° F. The equation E = 0.4 (T—48) fits Helgesen's field data quite well, and at 80° F this equation predicts 13 eggs per female per day, which agrees with laboratory experience (Wellso, per— sonal communication). Helgesen suggests his cages kept direct solar radiation off the beetles and thereby suppressed the oviposition rate. Mortality rates for the immature stages have been much studied, but the results are normally presented as total mortality for a given life stage during the entire season. Although this is a good start, more insight could be gained from a study which treated mortality per day as it changes during the year. Recent work in Canada has used the latter approach, but data analysis is incomplete to date (Miller, Gage, and Haynes; in press). The most complete study is that of Helgesen and Haynes (in press) which showed density-dependent mortality in the lst instar in oats and the 4th instar in both wheat and oats; all other mor- talities are density—independent (Table l). 3 4 TABLE l.--Mortality rates for the immature stages of the cereal leaf beetle (adapted from Helgesen and Haynes, in press) % mortality* Life stage Wheat Oats Egg 10 10 Larva lst instar 35 46D—85 2nd instar 3O 30 3rd instar 4O 40 4th instar 34D—31 28D—18 Pupa 30 30 *D is the common logarithm of the total egg input per ft2 for the year. Castro (1964) and Shade, et al. (1970) both attribute the coc- cinellid Coleomegilla maculata (DeGeer) with considerable predation on cereal leaf beetle eggs; Helgesen and Haynes (in press) assumed no pre- dation occurred. Consequently there is disagreement on the distribution of mortality among the life stages (Table 2), but the total mortality measured for the combined immature stages is quite comparable. Yun (1967) reported field data collected by Ruppel that indicated higher pupal mortality than the other two authors, but lower larval mortality; the combined mortality for all of the immature stages is quite compa- rable to the other authors. Overwintering mortality has been reported to range from a low of 25% in straw stubble (Burger, personal communication) up to 100% in 5 exposed sites 4 feet or more above ground level (Castro et al., 1965). An average value under good conditions seems to be 50%; snow cover and winter temperatures are the major factors affecting this value. TABLE 2.-—Comparison of mortalities measured by several authors at a common density of about 200 eggs per square foot in oats Helgesen Shade Ruppel Life stage and Haynes et al. (in Yun, 1967) Egg 10 56 50 Larva 82 65 34 lst instar 21 -- -— 2nd instar 30 -- —- 3rd instar 4O -— —- 4th instar 46 -- -— Pupa 30 48 91 Egg - larva 84 85 67 Egg - pupa 89 92 97 Dickler (1968) reported that full egg development under constant temperatures required 12 days at 60°, 5.5 days at 70°, and 5 days at 80° F. Yun (1967) gave the equation for % development per day (R) as R = .5473 T - 26.864 where T is temperature in °F. Dickler's rate is 1.4 times faster than Yun's. Dickler also found that fluctuating temperatures resulted in more rapid development; for example, eggs hatch sooner when temperature 6 oscillates from 60° to 80° on a 24 hour cycle than they do at a con- stant 70°, but he was unable to quantify the magnitude of this effect. Yun (1967) also gave the equation R = .2866T - 13.0896 for the % development per day for the larval stage; Helgesen and Haynes (in press) show that each of the 4 larval instars has a comparable developmental time, but their rate (Table 3) is 1.7 times faster than Yun's. TABLE 3.--Days required for each instar to complete larval development at 60°, 70°, and 80° F (from Helgesen and Haynes, in press) Temperature (°F) Larval instar 60 70 8O lst 3.81 2.55 1.86 2nd 5.33 2.12 1.71 3rd 3.00 1.87 1.44 4th 3.59 2.00 1.36 Total 16.24 8.53 5.91 % development per day 6.2 11.7 16.9 Field observations reported by Haynes (personal communication) indicate that temperature is not the only important factor affecting the developmental rate of 4th instar larval. If food is scarce, a nearly developed larva can apparently pupate early; also if the ground is hard and dry, pupation seems to be delayed a bit until it rains. 7 Pupal developmental rate as given by Yun (1967) is R = .2628T - 13.4378. Castro (1964) says ll-l4 days at 80°F, which agrees well with Yun's results. The cereal leaf beetle overwinters in almost any well protected site, usually at or near ground level. Castro (1964) found them under the bark of trees and in logs, in folded leaves, in straws on the ground, in the base of leaf sheaths and ears of standing corn, in baled hay, inside farm structures, in kindling wood, and even inside beehives. He found them in field margins, in the borders of wood lots, and deep within wood lots. Burger (personal communication) has found relatively high den- sities overwintering in wheat and oats stubble, especially where the stubble field borders on a woodlot or dense fence row to the north and/or east. Manley (personal communication) has found them consistently at about one per square yard in leaf litter deep within woodlots. Similar overwintering patterns have been reported for other chrysomelid beetles. Dominick (1939, 1971) and Dominick and Wene (1941) report that the Tobacco Flea Beetle is found "along the edge of wooded areas adjoining tobacco fields, along hedgerows, and in grass- lands"; many also hibernate in the soil around the remains of tobacco stalks. Emergence from overwintering sites for that species in the spring is spread over a period of 4 to 6 weeks with no well defined peak. The literature on mathematical models in population dynamics is very extensive. Two quite distinct approaches have been used: 1) keep the model simple and mathematically tractable at the possible expense 8 of realism, and 2) make the model realistic even though computer simula— tion may be the only method of analysis able to handle the resulting complexity. The distinction between these two schools of thought is well described by Liegh (1968). Pielou (1969) reviewed the use of simple models in ecology, and much of what follows comes from her book. All living organisms tend to expand their populations at a rate proportional to their current size, i.e., QEEED= r Nt, where Nt = # individuals at time t and r = intrinsic rate of natural increase. Such exponential growth can only occur at low densities where competition and natural enemies do not exert a density dependent effect. A more realistic equation that applies to a broader range of population densities is the so called Verhulst-Pearl logistic equation: §a%£-= r Nt (EFi—EE), where K is the carrying capacity or upper limit of Nt. When K is much larger than Nt, this equation predicts the same rate of buildup as does the exponential, but as Nt approaches K the rate of buildup is suppressed so that Nt approaches K asymtotically from below. Considerable effort has been expended searching for experi— mental verification of this equation; good fits have been reported by Gause (1934) for Paramecium, by Lotka (1925) for DroSophila and a bacteria colony, and by Odum (1963) for a yeast culture. But Smith (1963) with Daphnia experimentally measured %g%-and found a non-linear relationship between it and Nt, in contrast to the linear prediction of this model. Pielou (1969) states that this may well be the rule rather than the exception; and that the curve fitting methods of Gause, Lotka, and Odum were simply not sensitive enough to detect the deviations from 9 the model. Her argument is that the Verhulst—Pearl equation is but one of many possibilities for generating a sigmoid curve; hence when an ecologist finds his population grows sigmoidally, he cannot directly conclude that it fits the Verhulst—Pearl model. Leslie (1945, 1948) proposed using matrix equations to handle age specific birth and death rates and changing age distributions. A considerable body of literature has grown around this approach (see Caswell, 1971) and some authors (e.g., Lefkovitch, 1965) have experi— mentally verified their models. All of the above models were at first simply deterministic, that is only central tendencies are predicted. Bailey (1964) and Pielou (1969) treat the stoclastic generalizations, which include in each case a total accounting of probabilities. Biologists have made very little use of these stochastic methods, largely because of their mathematical complexity. Watt (1961b, 1968, 1970) advocates the use of complex computer simulation models in population dynamics; he has made several attempts to develop models with sufficient realism that management strategies might be tested on his models rather than on natural populations (Watt; 1961a, 1963, 1964). Most of his work, it seems, results in hypothetical examples of control strategies rather than actual evalua- tion of feasible stratigies; but even this is better than nothing, for his work has given recent stimulus to much concern with the complex model approach. Two examples are: l) the occurrence of 3 papers on "modeling insect population systems" in Forest Insect Population Dynamics (1969), and 2) the existence of a large National Science Foundation pro— ject at MSU on "Design and management of environmental systems" (GI-20). DESCRIPTION OF THE MODEL The computer simulation model developed for this thesis uses age specific oviposition, survival, and developmental rates and difference equations to predict daily population densities in each age class from a known number of overwintering adults as a starting point. A con- siderable number of simplifying assumptions are made regarding the effect of the environment on the beetles populations dynamics; these will be considered in detail after the model is presented. The mathematical form and logical structure of the model are described in this section; the FORTRAN version and a sample output are included in the appendix. The components of this model (Figure 1) are the life stages of the cereal leaf beetle and the internal structure of each component (Figure 2) is simply an accounting of transfer-in minus transfer-out. These transfer rates are computed from the above mentioned rates as reported in the literature. In the case of survival and movement of the adult beetles there were essentially no useable data in the literature, so observations were made in the field to gather the information. The detailed results of these observations are reported later in this thesis, but relevant facts are used as needed in this section. Emergence of spring adults from overwintering quarters is modeled as an impulse at 100 degree—days accumulated above a base of 48° F; 10 II SPRING ADULT I D = Death M = Maturation D OVIPOSITION EM ERGENCE '1? f OVER RING f :3 P am ‘y \ .L DIAPAUSE T O gogegogogego > C3 SUMMER ADULT Figure I. Components of the Cereal Leaf Beetle POpuIation Dynamics Model. OVIPOSITION DEATH b) INPUT 4. --.. MATURATION DEATH Figure 2. Generalized Internal Structure of a) the Spring Adult Component, and b) All Other Life Stage Components. 13 this corresponds to the time when 50% have emerged and is easier to model than a distributed emergence. After a feeding and host finding period of 10 days, oviposition begins at a rate of N eggs per female per day; N = 0.4 T - 18, where T is the maximum daily temperature in °F. Death of the spring adults progresses at a constant 0.7% per degree day. Egg developmental rates reported by Dickler (in Helgesen, 1969) were consistantly higher than reported by Yun (1967). A compromise value used here is R = 0.65 T — 31.6, where T is mean daily temperature in °F and R is % development per day. Egg mortality was used as 10% by Helegesen in his computations of larval mortality; however, his data averaged 14%. Other authors indicate that 20% to 50% is more reasonable, consequently Helgesen's estimate of lst instar mortality most likely contains considerable egg mortality. Since I will be using his estimates of larval mortality, I must also use his estimate of egg mortality as the two are computationally inter-related. So in this model egg mortality is a constant 10% during the life stage, although this will cause the model to predict overly large lst instar densities. Larval developmental rates reported by Helgesen were considerably greater than reported by Yun. A compromise value used here is R = 0.42T - 19.7, where T and R have the same units as for eggs above. Helgesen (1969) and Castro (1964) show that in the laboratory each of the 4 larval instars have nearly identical developmental rates; in this model they are taken as exactly equal for all temperatures. Larval mortali- ties are used exactly as given in Helgesen and Haynes (1971); total egg density determines the larval mortality rate, and no changes in 14 rate that may occur as time progresses are considered. Only in the first and fourth instars is there a difference in mortality between wheat and oats; for the first instar in wheat it is a constant 35% and in oats it follows M = 46 D - 85, where M is % mortality and D is the common log of egg input per ftz. Second instar mortality is a constant 30%, and third instar a constant 40%. Fourth instar mortality is given by M = 34 D -31 in wheat and M = 28D — 18 in oats. Pupal developmental rates are taken from Yun to be R = 0.26 T - 13.4. Pupal mortality is a constant 30%. Summer adults feed for 2 weeks during which mortality is 0.70% per degree day. They then diapause to overwinter and a constant 50% survive to emerge the following spring. Dispersal of the adult beetles and between field variance are included in the simulation model by considering a fairly large (but not precisely defined) geographic area containing both wheat and oats. The number of overwintering beetles, the acreages of wheat and of oats, and the portion of the population preferring wheat are set at the be— ginning of each computer run. The acreage of wheat is divided into three parts: 25% will contain a low density of beetles, 50% will contain a medium density-— 9 times higher than the low density, and 25% will contain a high den— sity—-9 times higher than the medium density. The oats acreage is similarly divided. This method of distributing the beetles results in the variance in density being related to the mean regional density according to the relationship reported by Ruesink and Haynes (in prepa— ration): 82 = 1.77 E 1'93. This corresponds to the observed 15 relationship between the number of individuals per unit area and the combined within and between field variance. When the adults first emerge from overwintering, the prescribed portion goes immediately into wheat, distributed among the low, medium, and high density parts as described above. The remaining adults are held aside until May 18 when they are distributed among the oats acre- age as described. On May 28 all adult beetles still alive in wheat move to oats; on June 15 all adult beetles still alive leave oats. Oviposition and mortalities operate independently in each of the 6 "fields". Hence in most cases larval mortality will be considerably greater in the high density "fields" than in the others. The model as described above actually obscures many causal mechinisms that affect the population dynamics of the cereal leaf beetle. Any mortality caused by weather, disease, predation, or para- sitism is considered background noise and simply contributes to the error of prediction. Of course, the data used in developing the model was gathered before any of the parasites specifically imported to com- bat the beetle had become an important factor, so their effect is excluded. The method of measuring and computing mortality used by Helgesen actually measures roughly from the midpoints of adjacent life stages rather than from their beginnings or ends. If the majority of the mortality occurs late in a life stage, then this method is very accept- able. For the cereal leaf beetle it is assumed in this model that all mortality occurs at the instant molting begins. This seems to be a reasonably workable assumption. 16 The developmental times are assumed to be precise measures with no variance term. This simplifies the simulation and should have very little effect on the analysis, unless the variance is very large. ANALYSIS OF THE MODEL Refining the model Before the model can be accepted as an accurate description of the population dynamics of the cereal leaf beetle, it must be tested against certain criteria which are predicated by our knowledge of the beetle. The most important of these criteria are as follows: 1. The average female beetle emerging from overwintering lays from 50 to 200 eggs in her lifetime. This range is not precisely known, but Helgesen (1969) reports in his thesis that beetles collected in early spring and held in field cages averaged 56 eggs per female; some eggs had been laid before he removed these beetles from the field to his cages, so this is an underestimate. An upper limit can be obtained by comparing total egg input per square foot to the starting density of adult beetles in the spring. Helgesen (1969) also reported 1,100 eggs at Gull Lake in 1969; unpublished results of adult densities show the peak density when adults first moved into oats was about 8 beetles per square foot. This gives an upper limit of 275 eggs per female if no adult mortality had occurred between emergence from overwintering and moving into oats. Any mortality during this interval would reduce this upper limit. Another source of data is Yun (1967), who reviewed the European literature and found 4 citations reporting from 50 to 150 eggs per female per season. 17 18 2. The generation index, the ratio of p0pulation size in one year to the population size the previous year, should be between 3 and 18 at low beetle densities, and should decrease to 1.0 as beetle den— sity increases to 5-10 spring adults per square foot. This criterion comes from the Cooperative Cereal Leaf Beetle Sweepnet Survey and from field plot data. 3. Adults should emerge from overwintering sites in late April, peak larval density should occur in mid-June, and summer adults should emerge from the soil about July 1. These dates represent the average conditions observed at Gull Lake from 1967 to 1971. The model as described in the previous section did not fully satisfy these three criteria. Two problem areas were recognized: l) fecundity was somewhat too low (58 eggs per female per season), and 2) summer adult mortality was far too high, 91%, resulting in a genera— tion index less than 1.0 even at low beetle densities. The two changes made to correct these problems are described below together with the reasoning used to arrive at the new parameter values. In the original model spring adult mortality was 0.7% per degree day. This value came from the 1971 regional survey and was first be- lieved to be the most accurate. But perhaps the value of 0.4% per degree day obtained in the 1970 cage study is better. One reason the 1971 value might be too big is if the insecticide killed more than the number of adults present in the sprayed fields at the time of applica— tion. This is expected if there is considerable between field movement of beetles. Also it appears from the data that mortality may be even lower very early in the spring. Hence in the revised version spring 19 adult mortality was used as 0.4% per degree day after May 18 (the date they move into oats) and half this before May 18. The second change was to reduce summer adult mortality to 0.05% per degree day. This particular value has no base in the data, but was chosen so that total mortality for the summer adults would be between 10 and 20%. There is a reasonable explanation why the cage study would accurately estimate spring adult mortality, yet very much overestimate summer adult mortality. Based on results presented later in this thesis, it seems that the cereal leaf beetle may overwinter fairly deep in the soil. If some of the beetles put into the cages for mortality studies completed their feed-out phase and began digging into the soil to overwinter, then they would not be found during the search for sur- viving beetles and consequently would be counted among the dead. These changes caused the simulation model to satisfy all three criteria. Further data are needed to determine if these changes were the prOper ones to make, but at least they appear reasonable and result in the model behaving in accordance with the presently known p0pulation dynamics of the cereal leaf beetle. Sensitivity of the model The next important consideration is the sensitivity of the model to small changes in parameter values. For example, egg mortality is set at .10; what would happen if that were changed to .20? The answers to this type of question are important because they tell us how accur- ately we should know these values and what the consequences of an error might be. 20 Perhaps the most important response variable in this model is the generation index (I), which is essentially predicted from the pro- duct of fecundity and the age specific survivals: I = F - SE - SI - 82 - S3 - S4 - Sp - SSA - SW' Of course, in the simulation model this I is computed as the ratio of adults surviving the winters of two subsequent years. But the sensi- tivity of I to changes in survival and fecundity can just as well be studied via this equation rather than the simulation. There are at least 3 different ways to view this sensitivity. The simplest is to consider the effect of a multiplicative change in a single variable. For example, suppose fecundity is doubled; the result is a doubling of I. If F is halved, I is halved. This relationship holds for every variable in the equation. A second approach is to consider the effect of changing survival by .10 for any given variable. Table 4 shows the results at an initial egg density of 1,000 per square foot in oats. Inspection of that table reveals that when survival is high, a .10 error has little effect on I; but when survival is low, a .10 error can cause important changes in the predicted I. This interpretation implies that this model should be more accurate at low densities than it is at high densities. A third approach to evaluating the sensitivity of I to the sur- vivals is to assume that each survival is known plus or minus its standard error; we can consider the combined effect of all 8 survival terms simultaneously by computing the standard error of I using the standard equation for the product of random variables as given in Yates (1953): 21 TABLE 4.——Effect of a .10 error in any particular survival estimate on the resulting deviation of generation index (I) from its accepted value of 1.08 Predicted I if survival is Age Observed Class (i) value (Si) Higher (Si + .10) Lower (Si - .10) Fecunditya 56 -- -- Egg .90 1.20 .96 lst instarb .47 1.31 .85 2nd instar .70 1.23 .93 3rd instar .60 1.26 .90 4th instarb .37 1.37 .79 Pupa .70 1.23 .93 Summer adult .84 1.21 .94 Overwintering .50 1.30 .86 aFecundity is 56 eggs per beetle or 112 eggs per female. b At a density of 1,000 eggs/sq. foot in oats. 22 SE (xy) -- x/or - SE >2 + (x-SE >2 If the standard error of the estimated survival is .10 for each of the 8 terms, and the standard error of fecundity is 10 eggs, then the cal— culated standard error for I using the values in Table 4 is .57; hence the generation index may lie anywhere between 0.51 and 1.65. At a lower density, I is computed to lie between 2.85 and 7.27. This sensitivity of the response variable I to the accuracy of the fecundity and survival estimates is an important consideration. If we wish to use this model to predict the absolute merit of a pr0posed management strategy, then we would probably want the model to predict I within :_20%. The above argument has just shown that fecundity within i;10 eggs per beetle and age specific survivals with 0.10 only result in an accuracy for I of about :_50%. So considerable accuracy is required of our ability to predict fecundity and survivals. If, on the other hand, we wish to use the model to evaluate the relative merits of several proposed management strategies, then less accuracy is required. In this case it is sufficient to know that strategy A will produce a larger generation index than strategy B, and therefore strategy B is preferred over A for the control of the cereal leaf beetle. A second response variable of particular interest is whgn certain points in the life cycle of the beetle occur. Specifically, when does the spring adult emerge from overwintering, when does peak larval den- sity occur, and when do the new summer adults emerge from the pupae? The date that adults emerge from overwintering is determined in the model by accumulating degree days above the base 48° F. beginning 23 April 1. The standard temperature distribution used in the model re- sults in spring adult emergence on April 26; if instead of 100 degree- days we had used 80 or 120, then emergence would have occurred on April 22 or 28 respectively. Since six days in April has very little effect on the resulting egg and larval population, an error of 20 degree days is of little importance. The developmental rate functions for egg, larva, and pupa can be interpreted as predicting the accumulated degree days above base 48 needed to complete development through those stages; they predict approximately 160,220, and 460 degree days respectively. The following discussion considers only the effects of prolonging developmental times beyond those used in the model, but an analogous argument can be made for the effects of reducing developmental times. If egg developmental time were 200 degree days instead of 160, this would cause egg hatch to occur 40 degree days later in the season. Consequently peak larval density and summer adult emergence would also occur 40 degree days later, which is 2 days at 68 degrees mean tempera- ture (the average for late June is about 70). So again a relatively large error in estimation of developmental times has very little con— sequence in the model. If larval developmental time were 260 degree days instead of 220, and egg developmental time were 160 as before, then egg hatch would occur at the normal time, but pupation and summer adult emergence would be about 2 days later than normal--exact1y the same end effect as pro- longation of egg developmental rate. However, since 2 extra days are Spent in the larval stage, the peak larval density would be somewhat higher than before and would occur about two days later than before. 24 If pupal developmental time were 500 degree days instead of 460, there would be no effect on eggs or larvae, but summer adult emergence would simply be postponed about 2 days. Any error of prediction of less than 7 days can probably be tolerated, with the possible exception of when the model is used to evaluate management strategies involving precise timing of insecticide applications. An error of 7 days would only occur if there were an across the board error of over 15% in the estimation of developmental rates. A third response variable, which is not important for the analyses included in this thesis but is very important in age specific mortality studies and in any work with predation or parasitism, is the size or density at any point in time of each age class. In other words this response measures the model's ability to generate total incidence curves. Since this response is not important in the context of this thesis, it will not be studied in detail, but let it suffice to say that considerable discrepancy is anticipated between observational data and model predictions. There are three basic reasons for this: 1) a 10% error in developmental time will result in a 10% error in the height of the generated incidence curve, 2) since all age specific mortalities are applied at the end of each life stage, the generated incidence curves will always be somewhat higher than they should, and 3) the model does not consider any changes in survival that may occur as time progresses; so if larval survival in May is better than in late June, the observed incidence curves will be peak earlier than predicted and have a shorter late season tail. 25 Predictions and interpretations The items covered in this section differ from those of the pre- ceding section in a rather subtle fashion. That section covered the response of the model to changes in certain variables which have ob— served values--values presumed to be constant over a wide range of space and time. This section considers the model's response to changes in parameters which vary considerably in space and time. First I should explain the method of analysis used in this sec- tion. Four different parameters are studied; each one is considered independently of the others. Except for the one being varied in each case, the others are held constant at what I refer to as its "standard" value. The standards are as follows: 1. Population size (CLBN in the program)——l,000,000 spring adults just ready to emerge from overwintering sites. 2. Portion of the adult pOpulation preferring wheat (RATIO in the pro— gram)--.10; that is, 90% of them go into oats the 18th of May. 3. Acreage of wheat and of oats available to the beetles-—100 acres each. 4. Temperature distribution--roughly fits the 30 year mean for Gull Lake and assumes the daily high is always 10° greater than the mean. It should be emphasized in light of the sensitivity analysis in the preceding section that the following interpretations and predictions may lack accuracy in regards to the absolute value of the generation index predicted, but they should be very accurate relative to one another. Initial density: The starting size of the spring adult population was varied from 10,000 to 100,000,000 which corresponds to .001 and 10 per 26 square foot of grain respectively. The generation index (I) was at a maximum of 6.0 at the lowest density and stayed there until the density reached about .02; I then began declining, until about 7 beetles per square foot it reached 1.0 (Figure 3a). I continued to decrease, getting close to zero as the initial density became very large. This is exactly the behavior expected. Helgesen (1969) accurately described this as a density-dependent feedback system, the requisite for population regulation. The biological interpretation is that when- ever initial density is less than 7 beetles per square foot, the popu- lation will increase; whenever it is greater than this, it will decrease. Hence regardless of what the initial density is, after many generations the density will be close to 7 per square foot. One point needs emphasis here: not too much importance should be attached to the value 7 per square foot. There are two reasons I say this. First is that the sensitivity analysis showed the model could be about 50% off in predicting the generation index; reconsideration of Figure 3a knowing this shows that I may reach 1.0 at any density between 3 and 18 per square foot. Second is that this 7 is average for the area. Some fields always have more, others less, than the average due to between field variance. The important interpretations here are that population size is density regulated and that density-dependent mor- tality begins to exhibit its effect on the generation index when regional density is still about 400 times less than the carrying capacity. Portion preferring wheat: At first it may seem that this parameter should be a constant, somehow determined by the genetic makeup of the beetle population. Or maybe it should even be a constant 0.0, because Figure 3. 26a Predicted Influence on the Generation Index of Varying a) the Initial Population Density (0) of Spring Adults, b) the Portion (P) of the Population Preferring Oats over Wheat, and c) the Portion (R) of the Small Grain Acreage that is Planted to Oats. INDEX GENERATION 27 O .OOI .OI l I I0 IOO ADULTS PER SQUARE FOOT (D) { J A --o--o-----o--- W ‘o —o— P = .5 —o—P=.9 D = .I 7 ' 1 L O 5 LG PORTION PLANTED TO OATS (R) Figure 3. 28 all of the beetles actually prefer oats. But, in fact, this parameter is quite variable and quite important, considerably more so than was thought by most researchers until now. The response of the generation index to changes in beetle pref— erence is quite small (Figure 3b), only ranging from 5.85 to 4.62 as preference moves from all in wheat to all in oats. And most of this difference is due to the fact that beetles which prefer oats effectively lose part of their potential egg output before May 18 because no host is available for oviposition. More important is the movement of the spring adult among host plants, a factor not reflected in the reSponse of the generation index. This consideration is of extreme importance if one considers directing control strategies against the adult to prevent oviposition. Available host acreages: There are two separate issues regarding host acreages: 1) what total acreage of small grains is available in the region, and 2) what portion of that acreage is spring grains. The first, total available acreage, affects density and consequently affects generation index as shown in Figure 3a. The second issue is a bit more interesting. The ideal habitat, from the point of cereal leaf beetle population buildup, is one that contains both winter and spring grains (Figure 3c). Both pure oats and pure wheat cultures suppress the generation index (I), primarily because fewer eggs are laid when oviposition is restricted to one crop or the other. But I is also suppressed when 99% of the small grain available is one crop and only 1% is the other. In this case the reason is that when so many adults concentrate in a small area, 29 the resulting high egg and larval density causes an increased larval mortality. It is interesting to note the interaction between host crop rela- tive acreages and the beetle's inherent preference for oats. If 90% of the beetles prefer oats, 1 peaks at 5.1 when 90% of the small grain acreage is in oats. If 50% of the beetles prefer oats, then I peaks at 5.45 when 50 to 60% of the small grain acreage is in oats. So it seems that optimal rate of population growth occurs when the proportion of small grain acreage planted to oats corresponds to the proportion of cereal leaf beetles preferring oats. Temperature distribution: Temperature affects the beetle population three ways in this model: 1) oviposition rate is directly proportional to maximum daily temperature, 2) adult survival is inversely proportional to mean daily temperature, and 3) accumulated degree~days determines when events such as emergence from overwintering and peak larval den- sity occur. Five different temperature distributions were considered (Table 5), corresponding to the 30 year averages for five geographic locations within the current or anticipated range of the cereal leaf beetle. Lansing and Alpena were chosen because we know something about the beetle's dynamics in these areas. ConSequently, the predictions of the model for these two areas can be evaluated with respect to accuracy. The other three areas, Bismark, N. D., Ottawa, Ont., and Louisville, Kty., lie on the periphery of the beetle's current range. Predictions of the beetle's dynamics in these areas will be useful for timing detection surveys and for anticipating the rate of build—ups. 30 TABLE 5.--Average daily temperatures for April through August at selected locations in the potential range of the cereal leaf beetle Location April May June July August Bismark, Ha -- -- -- 86 -- N. Dakota Lb —- -- -- 58 -- MC 44 56 65 72 69 Ottawa, H 50 65 75 80 78 Ontario L 31 43 53 58 55 M 40 54 64 69 66 Louisville, H 66 76 85 89 87 Kentucky L 43 57 62 67 64 M 54 66 74 78 76 Alpena, H 52 66 76 80 79 Michigan L 29 39 49 53 53 M 40 52 62 66 66 Lansing, H 56 68 78 83 82 Michigan L 36 46 56 60 59 M 46 57 67 72 70 8H is the average daily high temperature. bL is the average daily low temperature. CM is the average daily mean temperature. 31 Table 6 presents the summarized results of this analysis. The dates for Lansing and Alpena are as expected, so the other dates are probably close to what would actually occur. The expected north-south gradient is quite evident, except that Bismark compares very closely to Lansing rather than to other more northerly sites; but Table 5 shows that Bismark's temperature distribution is also very much like Lansing's, so comparable dates should be expected. TABLE 6.——Generation index (I), fecundity (F), summer adult survival (S), and the dates for spring adult emergence (SP), peak larval density (PL), and summer adult emergence (SU) as pre- dicted for 5 locations Date Location I F S S P P L S U Bismark, N.D. 6.1 147 .84 May 3 June 13 July 7 Ottawa, Ont. 6.6 154 .87 May 9 June 16 July 14 Louisville, Kty. 3.5 84 .81 Apr 13 May 22 June 15 Alpena, Mich. 6.9 156 .90 May 10 June 21 July 22 Lansing, Mich. 5.4 128 .85 Apr 29 June 13 July 7 A north-south gradient is also evident for generation index, fecundity, and survival of summer adults. Fecundity is highest in the north where, l) cooler mean daily temperatures cause the ovipositing adults to live longer, and 2) at the same time the day-night temperature fluctuation is greater, so for a given mean temperature the daily maxi- mum is greater than in the south; daily oviposition is determined by the maximum temperature, so more eggs result. Summer adult survival 32 is highest, as expected, in the coolest climate. The combination of increased fecundity and increased survival of summer adults causes the generation index to be highest in areas of cool mean temperatures with large day-night fluctuations. There are at least three factors not considered that may have important effects on Table 6. One is the influence of relative acre- ages of winter and spring grains in each region. The present computa- tions were made assuming equal acreages of each, and small deviations from this should have little effect. But in fact North Dakota has essentially no winter grains and Kentucky has essentially no spring grain. Just how the beetle will adapt to these conditions, and what portion of the resulting resident population will prefer spring grain over winter grain remains unknown. The second factor involves the causal relationships affecting oviposition rates and adult mortality. If these rates are not related to temperature, or if they are related in some fashion considerably different than currently modeled, then fecundity, summer adult survival, and generation index as reported might be far from accurate. The third factor affects Kentucky and other warm climate areas; emergence from overwintering sites is modeled as occurring at 100 degree-days accumulated after April 1. In the north this works well, but in Kentucky March is warm enough that in fact the beetles may emerge before April even begins. Since this possibility was not con— sidered in construction Table 6, the dates there for Kentucky must be biased toward lateness. If spring adult emergence occurs April 1 rather than April 13, then 1) fecundity would be greater (maybe about 100), 33 2) generation index would be greater (maybe about 4.2), and 3) both peak larval density and summer adult emergence would occur earlier (maybe 7 days earlier). Daily temperature fluctuations: If daily temperatures followed the smooth curves Obtained from the 30 year means, then the total incidence curves for the immature stages of the beetle would also be very smooth. Instead we have warm and cool days intermixed, which cause egg input and larval developmental times to be irregular. These irregularities in turn cause the total incidence curves to be quite rough. For ex— ample, Figure 4 shows the total incidence curves for larvae that the model generated using the temperature distributions at Lansing and Alpena, Michigan. Especially notice the sharp drop in density from day 79 to day 80 on the Alpena curve; in one day density dropped over 35%, then climbed back to nearly its original level before the normal late season decline began. The following sections of this thesis pertain to the field work that was done to support the model. 34 20' 0) IO r i.— C) C) LL. LLI 0C <1 :3 C) 1 1 . IC3 (0 5/30 6/l5 6/30 7/l5 DE 3 30 F 2% b) :> a: j 20 r IO " O l i 5/30 6/I5 6/30 7/l5 DATE Figure 4. Two Examples of Predicted Day to Day Fluctuations in Larval Densities: a) for Lansing, and b) for Alpena, Michigan. 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 the Kellogg Biological Station, hence is under control of Michigan State University. 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 1,315 acres under cultivation was distributed among about 300 fields ranging in size from 0.8 acres to 35.8 acres. The remaining 527 acres was subdivided as follows: woods, 249 acres; fence rows, 13 acres; roadsides, 27 acres; weeds, 25 acres; and others, 213 acres. The final category contains such things as lakes, roads, buildings, and lawns. None of these were sampled as they were considered unavailable to the cereal leaf beetle as habitat. Overwintering sites An extensive search was conducted to find the preferred over- wintering sites of the cereal leaf beetle. Several methods were used, the principal one being to dig up a sample (roughly 3 square feet and 3 inches deep) of earth including all plants above that area and run it through a cotton gin trash mill. This machine was acquired from the Plant Pest Control division of the U.S. Department of Agriculture; they 35 36 had designed and used it to survey for pink bollworm larvae in the trash left from ginning cotton. The machine consists of two revolving screen cylinders which sift the material of the sample into three parts according to particle size. When the soil was loose and dry, this machine efficiently sepa— rated 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. Especially designed emergence traps were used in the spring of 1971 to sample the number of beetles emerging from overwintering sites. These traps covered a square yard of ground surface and caught emerging insects in a pan of glycol when they reached the tOp of the screen sides. Beetles were also found in their overwintering sites by direct Observation. 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 a field cage technique. The cages used were 6-1/2 feet square and 6 feet high with plastic screening for the sides and top and with a zipper door in one side; the bottoms were open so the cages could be placed over the host plants in the field. Two cages were used for spring adults; the first three weeks they were in wheat, the last two in oats. Four cages were used for summer adults; two in oats and two in corn for the entire three weeks. In every case when a cage was first set up in a new 37 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 1/2 hour and then came back and re- peated the search. Normally the second search caught about one-tenth 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 in- side 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. Hence the % mortality over the sample interval was found from this equation: I - D - R M — 100 x I _ D , where I = no. put into cage, D = no. found dead in jar, and R = no. removed 6-8 days later. Population survey Each grain field in the study area was sampled to determine the number of beetles in that field; the sum from all fields estimated the number in the region. Every field was sampled at regular intervals, normally twice a week, to detect any change in beetle population. Mor- tality of adults was computed by comparing the population on successive sample dates. 38 The sampling technique used varied with crop height. 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: l) throwing a 12 inch garden stake at least 10 feet from where one stood, 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 technique 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 (D) by the equation given in Ruesink and Haynes (in preparation): D = c (0.20 + 10K). where K = -.06 + .02 H - .017 (T + 10S) + .661 loglO(W + 1), H = grain height (inches), T = temperature (°F), 8 = solar radiation (cal/cmZ/sec), and W = wind (mph). Most of the needed weather data were available from our own weather station set up within the study area; however, some data came from U.S. Weather Bureau records for Jackson, Michigan. RESULTS Overwintering sites The gin mill samples taken August 11-18, 1969 found beetles in nearly every habitat surveyed (Table 7); the total estimate of 10,351,000 beetles in the 4 mi2 region is reasonably close to the 5 to 10 million expected based on population levels the following spring. By late September and early October, however, less than 1 million could be accounted for by using the gin mill, and by early November only 430,000 were accounted for (Tables 8 and 9). Hence it appears that actual overwintering does not occur in the surface litter nor in the tap 3 inches of soil. Ground litter samples collected March 20, 1970 were held at 65° F in the laboratory; 4 beetles emerged from 72 square feet of straw stubble, 1 from 4 square feet of fence row litter, and none from 4 square feet of an old hay windrow. This accounts for 2,420 per acre in stubble, or 636,000 in the 4 square miles, perhaps 10% at the most of those present in the 4 square miles; the fence rows may account for another 2%, according to these meager data. Straw stubble was again checked in the late summer and fall of 1970. Beetles were found in 6 of the 9 fields surveyed at an average density of 0.14 per square foot, or 6098 per acre, which is 1.7 million in the 4 mi2 area. Three yd2 litter samples from fence rows gave 15 beetles, or 0.3 million in the 4 mi2 area. 39 40 TABLE 7.-—August 11-18, 1969 gin mill finds by habitat at Gull Lake. Each sample consisted of 3 square feet # # # CLB/ # Total # Habitat CLB Samples acre acres CLB (1,000's) Croplands Idle 3 7 6,222 529 3,291 Grain Stubble 11 21 7,606 263 2,000 Alfalfa l 9 1,613 216 348 Corn 2 6 4,840 267 1,292 Subtotal 17 43 -- 1,3038 6,931 Non—croplands Woods 7 12 8,470 250 2,118 Fence rows 13 6 31,460 13 409 Roadsides 37 21 25,583 27 691 Weeds 5 9 8,066 25 202 Subtotal 62 48 -— 5273 3,420 Total 79 91 -- 1,8308 10,351 aIncludes acreage not sampled, but CLB density assumed zero. 41 TABLE 8.—-Late September and early October 1969 gin mill finds by habitat at Gull Lake # # # CLB/ # Total # Habitat CLB Samples acre acres CLB (1,000's) Croplands Idle 1 18 807 529 427 Grain stubble 0 30 0 263 0 Alfalfa 0 26 0 216 0 Corn 1 45 323 267 86 Subtotal 2 119 -- 1,3038 513 Non—croplands Woods 1 9 1,613 250 403 Fence rows 4 23 2,525 13 33 Roadsides 0 15 O 27 0 Weeds 0 20 0 25 0 Subtotal 5 67 -— 527a 436 Total 7 186 -— 1,8303 949 aIncludes acreage not sampled, where CLB density assumed zero. 42 TABLE 9.-—October 23 - November 6, 1969 gin mill finds by habitat at Gull Lake # # # CLB/ # Total # Habitat CLB Samples acre acres CLB (1,000's) Croplands Idle 0 9 0 529 0 Grain stubble 0 3 0 263 0 Alfalfa O 12 O 216 0 Corn 2 18 1,612 267 430 Subtotal 2 42 -— 1,3038 430 Non-croplands Woods 0 3 0 250 0 Fence rows 0 0 (0) Est. l3 (0) Est. Roadsides 0 4 0 27 0 Weeds 0 3 0 25 0 Subtotal 0 10 -- 5278 0 Total 2 52 -- 1,8308 430 aIncludes acreage not sampled, where CLB density assumed zero. 43 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 was my first sample from such a habitat and indicated that signifi- cant numbers of beetles may overwinter in micro-habitats which are very hard to quantify. In early April of 1971 some more searching was done to find out if large numbers of beetles overwintered in such cracks in logs. Of the 162 beetles found in 4 old fence posts, a decaying stump, and under wild grape bark, only 9 were alive (Table 10). 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, so overwintering mortality in above ground exposed habitats is estimated at 94%. Of the 32 emergence traps set out in the spring of 1971, 18 caught one or more beetles, 2 were nonfunctional, and 12 others caught nothing. The detailed catches are listed in Table 11, while Table 12 summarizes catch by habitat. These 32 traps were not placed at random in the environment, but were placed in sites where overwintering beetles were expected based on previous gin mill work. Still, by proper strati- fication of the environment, an estimate of the beetle population can be made as follows: consider 4 beetles in 7 yd2 as the average density for all crOpland, woods interiors, and the so called weeds. Estimating 20% of the woods acreage as edge and 80% as interior gives us 1528 acres at the above density, or 4,226,000 beetles. The high density area is made up of woods edges, fence rows, and road sides. In this area 109 beetles 44 TABLE lO.——Overwintering mortality in above ground habitats; Gull Lake 1970 - 1971 # CLB found in April 1971 Habitat # Dead # Alive Fence post #1 8 0 Fence post #2 0 0 Fence post #3 139 7 Fence post #4 l l Decaying stump 0 0 Under wild grape bark 5 0 Totals 153 9 153 % mortality = x 100% = 94.5% 153 + 9 45 N o o o o o o o o o o o o o o o o o o N o o NNNo NHNH «N H o o o o o o o o o o o o o o o o o o H o o NNNo NHNo NN a o o o o o o o o H o o H o N o o o o o o o NNNo NHNS NN oH o o o o H o o o m o N o H N N o o o o o o NNNo NHNo ON o o o o o o o o N o o o o o o H o o o H o o NNNo NHNo NH H - - -T - - - - - - o o o o o o o o o H o o HNNN NNo oH H - - - -T - -T - - - o o o o o o o o o o o H HNNm NNo NH N - - - - T- - - - T- o o o o o o o o o N o o HNNm NNS a mo 0 o o o o o o H o o o N o o o o o 0 mm N a mHNN NNo N N - - - - T- - - - - -T - - - - - - - o H o H NNNS NNo m m o o o o o o o H o o o N o o o o o o o o o NNNo NNo o H -T - TT - - - - - - o o o o o o o o o H o o HNNH NNS N N T- TT - -T T- - T- - - o o o o o o o o o H o H HNNm NNS N W l G nu l m NN NH EH NH w m H N NN HN NH 2 S N m om NN mN HN NH 3 m m m m T. \o \c \m \o \o \o \o \m \m \m \m \m \m \m \m \S \q \q \q \q \q m s "v D. a 9 a 0 ate No touuo Han muHSpm wafiumucwsnm>o pom mamuu munmwuoeo CH guano-I.HH mqm.HDHIII onLII1100~111199H1111xH11113 7o 1xHHII).XHhII). XOLII) XUI(I) XOHII). TEII).ODDII) 30 CONTINUE COMPUTE AND PRINT YEAR END RESULTS 75 CLBNY:(11-DQ):XHIN A}D:CLBH/(HHA+0AA)/4SSOOu D:=1.-XHIk/TAE HHITEL§11§131_§L§11§L9NY HRITEI611506) NHApOAA so HOITEI61.304) RATIO HRITE I§1_ 331) TE11I_AE, xHIN HQ!TE(61:309) AID HQITEI61.309) 923 O45 0561078 D8109 HRITE (61 31111P§161 _O_34H._O34H D34OL_ D340H1034OH 067HL DbZH_1DA7NH1 85 10670L 0610M. 0670H EHL. EHH.EHH E0L1EOH EOH FORMAT SIATEHENTS 301 FORMATI1H1o3nX.‘DAY BY DAY RESULTSa.20X.9DAY= 1 IS APRIL 1ST*/O4*n 9o 129x.15P2I G AOOLT OFNSITY~1§§H1HSUAMEH_HQHLI EHFRGE nest/c gssx,, 15R(1H*) 2X'r8(1H"/* DAY‘2‘6X;'WL*08X5*HH.O8XO‘HH'OBXO’0L:18xa‘OH0 1. ax.90H..EX) ~ TEMP 509) 302 FORHAT (I4 12F16 6oF5. 0 F6_.111. 303 FORIAT (1+1. 32x.9YEAR EHO SUHIARYo/vQSTARTEn IITH91F15 0,65PRING A 95 1DULTS 0/6X19AND GOTtoF15.OI9SPRING ADULTS FOR NEXT 11615 I 304 FORMAT( 9-P0HTION OF THE REETLES PREFERRING HHEAT*.F6.2) 306 FORVAT(o-AVA1LARLE ACREAGE THIS YEAR HAS -9.F10. 0.9 OF "HEAT: AND. 1/31X15111699_ 0F OATS. 9) 100 307 FOHIAT(.-TOTAL EGG IA.PUT HA391F18, 01* EGGS.9/9 TOTAL ADULT EMERfienw . 1CE HA891F12. 9.9 SUMMER ADULTS.*/t 6.22X.*'N0*o712. 05* ADULTS WENT 1 JNIO.OHE“HIHIEBING9IMM- 308 FORIAT(.-AVFRAGE INITIAL DENSITY ~9.F10. 4:0 PER SOUARE FOOT9) 309 FORMAT(9'IE“SITY INOEP EHOEHT HORTALITIES 99/1cx.9EGGS*oF7. 2/1OX0 105 11L21LEQLKILJXL1L§11£912£11XI9PUPAErLF6.2/10X;9SUH A09IF512/10X. 16H1NTFRQ, F“,?) 1 310 FORMATI9HFErsITY DEPENDENT RATES *9.10X.thEAT9.41X.*0AT59/I 1 27X2'L0H'1:Xn*MEDIUM#,5X'Ofllfi 1H6117X.9LCH9,2X.9MEPIUI9,5X.*HIGH9/12X.9L1 HURT. ..3r1o 2,1ox 3r10, 110 12/12X.*L4 NORTo5o3r10.2.10X13F10.2/ 11X.*tGG INPUT*03F1092010135F1 10:2) __ EHD SUBPOUTI*E STAOIA (RI RO.XH .RO STOR HURT) DIMENSIOJ,SIC”LLDDI_.__-_ REAL HURT “" 5 _ THIS ROUTINE UPDATES THE DENSITY 0F ALL IHHATURE STAGES CALL LAG (HIIHOUT13fUH1Hb) ‘ ‘ anFOUTo(1.-H3RT) _ X H =51: 811, 11913-- END ' _ ”" "“‘"_"‘“"“ "‘—" ‘ SHBROUTINF RA 'EX (R'SIG,RHI THIS IS A FOR? IYH ”CUT_I_NE USED TO_ GENEPATE AH AUTOCOPRELATFD SEOHJE CE OF NUIB 13. Z=RANF(-1) ‘5 RzRoRH9S15*TA~LIEIO o. 0.25.2c_ z) RITOHH ~v*~~v~w*—-~~WAHLHEH_AAW_911__111_,. END 73 SUBROUYINE ngA (DS4.D67.TE.XOUf.EI.X9 )'51118?ILX11§7’ 10 15 20 25 30 Si, 40 10 20 100 DIMFNSIQN_IEL1§11LXDU111811L18(10)oR051 ornewsxow 52(1oo>.53<100).34<1o COMMON/92/023;oas.oso.D73 ZERO our NECES§§RY MEMORY; on): )555(100).56(10c).s7<100) R12=0. R2330. R34eo. R4580. R56‘00 R67=0. R78=0. . _ DO 10 131110 XEJIISQ, , '_ DO 20 :=1:100 $2(!)=0. SS(!)30. 86(1)80. SS(!)IO. 56(1180, S7(I):0. CONTlNUE . Dn 100 1311187 COMpuTE FRACTIONAL DEVELOPMENT PER DAY FOR EACH LIFE ST‘GE Rfij?)=.65fjfjll:§lg§ Rfi(3)=1.65'TE(l}P78.8 Rn(4)=RD(3) 8212138013) RD(6)=RD(3) Rn<79:.26*TE(1)w13.4 UPDATE DENSLIjEE IN Efigflngfg STAGE X(I)IR78 XDUT111=R7figx9u1LL3 CALL STaDiA (R67,R78,XN(7).RD(7);S7;D78) CALL STADIA (R56.R67.x~(6>.RD(6).36.0671 CALL SIADIA_LB€§1856.XNLSILRD151185.056) CALL STADIA (R34,945.XN(4),RD(4).S4.D451 CALL STADIA (R23.R34.XN(3).RD(3).S3:D34) CALL STAQIA (R12.RZ3JXN(211RD(21;§2oDZBI R12=El END 10 000 10 SHBPOUTINE SUM AD (RIoXNINonDD) DIMENSIQE.SIQBLLQQILDDUL187>LR{1187) Do 10 1:1.100 STQR‘1)300 THtS‘fihUTTNE—ffitfdHS"TEE‘§E€YEE§”?§OF”PUPKE‘€HFE5EfiEE*iNTO“6VEEEXWIER¥“ iNG AND APPLIES A HOQTALITY WHICH MUST BE SET IN v-r xsvavt . 1 . Rn=7| XNIN3UQ On 20 1:1,167 XSUPV;.99QSgLflQD(I) 15 ffl_1§_§IQB(J):ST09(J)ixSURy_ 20 CaLL’LAGoDDHIEB7I5 1 FIHLI1R7).FIHMI187).EIHHI187)aEIOL(187).EIOHI1R7),EIOHC187) DIMENSION TE(187)oDDD(187) THIS ROUTINE MONITORS SPRING ADULT DENSITY AND MOVEMENT FETHEEV CROPS. IT ALSO COMPUTES DAILY EGG INPUT INTO EACH OF THE 6 FIFLDS, 10 15 20 25 30' 35 4o 45 50 55 10 0 00000000 11 12 00300 EUL‘O. ENM=O. EHH‘G. EOL=O. E0H30. EOH=0. DO 10 131.187 HDLIII=OI WDH(I)=00 HDHIII=00 OWLIII=01 ODHII)=09 ODH(I)300 EINL‘I)=00 EIHVIII=00 EINH‘I)=00 EIOLLII=UD EIOMII’BOO EIOHIIIIOu DO 20 I31116 SPRING ADULTS EHERGE FROM OVERWINTERING AT 100 OD BASE 48 IST IS THE DAY OVIPOSITION BEGINS. 10 DAYS AfiTER ADULIC"FIRsI snow; serwgfih EjELD VARIATION Is APPROXIMATED 9v 3 DaggIII CLA§§ESI see rexr rDR DETAILS. IPIDDJEEJADD.) GO 10 11 DD=DD+DDDIII Ir0~o .- f. ‘ f} ‘1‘ HHr‘civOrC . H O D 0' V. 0"... .vm ‘cocomc.c Iconoco.o ooooucuc.Aooooao.q1roomcooJo-Ianaocoaallaouooo.aliaoouoo-a aaoucha «domco. .deNdo. anoaq. flail. .om untauL. oocorcno oaocounc cocooo.o oogouo.o oogooo.o cocooo.o aooooo.c coeCDc.o aocmoD ~nouoo. mvmooo. on .n Iuaocg. cococu.o GOCOQOH , nocooo.o caboco.o ocgnca.o oogooo.o concog.c cocc,c.= oQNocu mvssco. “cage”. owal u u “III d .5» “c.o¢‘. omcho.o oupuucnc oououo.a.ncono.0.azzdcuono”anuoyoca.oanon990.9:Idocc»u.q snuanu unoQNQOq;s:-qocu. «a . fl .ne Anna... oc.c¢u.o oncogene oogooo.o oogo;o.o occouo.o oo.coo.o oaaoog.o coocuc.c ocunuu OAHnoo. Nooooo. “a ” .nn “0‘0... ooporo.a OUUOOCHaYYaucooo.o Ycouogowo ocooco.o ocoooo.c oomooo.o coccgc.c convuo. .ommoo. «Haven. cw ” .u. (c.0LD. rococo.o oonouoao DOLooo.o..cocoaoaOIIOOQDOOUO.Ioooaua.qtldooouo.aieooacuo.anuoaguuqu.nooaoa.a-;oocooa.q::m~:1 A .0. cococ.. nocouu.a nooocouo oocoo°.o oouoOOJO ocoooo.o ocuooo.o oouooo.a coacqc.c cocoau.o ooaooo.o oocooo.o .n ."w cocaUU. o.no¢o.° cacooeng oo‘ooo.o ccuoooqo accououo oo;ooo.o canooo.° coacuc.c ocoogng oocooo.o cocoon.o nu . .Im coccgg. nouooo.o eczema"; coaooo.ot ooueoo.oT:oococouo.xoccouo.a:noooooo.aTIaQQCJc.a;Tdaocouma oouooo.o.‘oocooo.o ,mmil . .fi. uo‘ouc. oopo‘o.o casemeno oooouo.o cagoco.° cocoo0uo ocuooo.c canoo°.a coococ.c oocooo.o oopooo.o oocooo.o «m . .hn AUDOID. nauomo.o ouoooonc oncooo.o cocogo.o cocooovo ooaoco.= soucoo.o coccsc.c occoguuo oococo.o oocuoo.o on a .vm cc(cct. oococo.c cccoocna o0&oco.o coacho.o occooouo neoooo.c oooooo.a -coccwc.c noocuumo ooooco.o ooooon.o on.a. . .mm Ac OLD. coacou.o unmounDC cabana.“ concuowo ogaOQONO cagooo.o ooouoo.a coucgc.c oooououo oopooo.o oococo.o g“ .n¢ “smog“. o¢v cccogc. .nc .cococc. «Hm soccer. a. «C(ocg. .mm “recur. acfi hereunto .vm .cAQL.. unpopu. .a¢ tugccu .cm cucocc. on uncccu. .nv co«o¢¢. .a' cccocv. mapouu.o ceoouo.c aocoao.o oocouo.o engagemo an.aoo.o oaauac.a coccac.c ooaooo.a nooooo.o oocooo.o a" pognuo.c cccoucnc oococo.o .oogogoao.|occocouo-Ioouoao.oIzouonno.cg:coccuc.c-IGOQOQUHO, socooo.o; wococn.o; («II raceco.o ccccce. concoc.o .oaoogo.o occoc°~o gauooo.o .ooooao.c coccqc.c oocoouna oocooo.o cacooo.o ma ancgyg.o ocoooo oocooo.o _oaooQowaIloococano roouooo.c 399000.; cooCA=.c nooooo.o aooooo.a oouooa.o .« coccco.o cannoo ¢opooo.m cauoto.o ocoooouo nogoo0.a,sooo”no.qrrccccmc.qTlduucgomo.3o°oooo.o -aocooo.o.-nnzl ogvuc°.o occooo oooooo.u ooccuo.o oogo;a.o ocgouo.o ooooaa.o caca;c.c “Doggone sooooo.o cococo.o Nu zucchu.c Tongue. co;omo.u .eogouodoxrmo¢05°roz.oo”ouo.ps-oooumo.psgcoacacucIImogcqo.oAxoooooo.o cocoou.o «H naho.u.o «coupe. O'.!.'. .0. N NVVNWO'mnvNNdmNo .00... 00800: comooo.o co c¢o.° ocgacoho GQJooo.o coacoo.c cacaac.o oosouomo accooo.o.aoonoon.o an noLccn.o couuocnc uopopo.o accepo.o ongogouo ocuooo.o 900090.” ooongc.c acooaoua cooooo.o oomouo.o o nmcuko.o romouo.c “abopo.o homage”ozqocwooo.olabomooa.QIIMppuoa.nltmoucgc”caImo“cup.©I.nouooo.o-:notooo.o a noccAo.o tongue»; mo‘ooo.u ooaofio.o occooo.o accoquo canopo.a coecac.c ooocuom aocooo.° oo:oon.o n ooccn°.o casemena ongooo.o Docho,o accouo.o aauooo.a cocoao.o coccgc.o acccuo o oooooo.o oanoon.o o OGCCID.O cooaoa.o cocoOo.o- poacLOau.zoooooououroamooouo ooaoaa.o coa:bc.o gogao o ooooco.o occoco.oklm moroco.o anaconda cocoog.° cocoQ o . cococo.° . .5. COAUAA. ocuaco.o cocooo.c cocom0.0 cocoa o ocuooouo oooooo.a cocooo.° ooccnc.a oooooono oocooo.o oooooo.o » .on cacaoc. oaoouo.a concocuo cowoco.o-:oou0p o apoooo°.o Toooooo.oaaooooo°.o .ouocdc.c oouoao.o--aouooo.a.sooooon.o.Im p o nogooo.o cocoou.o H > '9 fl . r. O - O O O O I’II I . ‘ I,II|II I..I.¢I OGQOQONO Ioaaooo.oyldoooao.a ccwcac.cnsdocaoo o ooodo°.o ,. ”O I. ‘0 I. 0 , o . o GOCOCOOCCCOCCDCOOCUOGCGCCCODO (a o o o . - . . ;; A -- - .cn cu‘occ.g oococa.o cocoao.a comooo.° coco o. occooo.oI uooouo.c ooguoo.c coca;a.c nogcoouo no to do :2 z: 43 no :0 40 13 x: 43 00chuoo¢ooc0000.00.00.0a.¢o.c.acoccnoouCooococccccctctoccc COOC4CCCCCCCDC¢COCCCGlace...CCOCCOCOOcoocccooco00000000000 uuzwcawzm chag qmzxam gunshonononhmooacadnmv«INduansomuhmmnomvoow'mno ' . ‘4 Do 0. 1 LU .— »hmwzmmiugaa‘ czmmmm hm“ 4~mn¢ m" ".>m > oo. o oco . o N“ ooo o coo Io ooo o ooo o ooo . o oo co.o co. co. no. oc. -I v o.a can oo o IoooIooo ooo acoII.co c.o IIdo coon o cooo.o . roo; o NHI 00-0000 00 000.0II0 .00 .0 .0 0000.0 0 0000. 0 000. 0 ”000 ..0 000I 000 0 .00 0 000 0 00 0 00 0I 000 .0.0 .0..00. 00 o.o ooo oo. do; oo. oo coo. o ooo. o oco. . .0: o . 0 o .0 o 00 o 00 I0 00 0. 00 0.0-.00 OOOIOOo OD uouII 6 Dc. 9 can“ a 000- o coo: d 000.0 0 000 0 000. 0 0000 .0! 000 qII 000 ..0 000 I o ooo.o.o coo o.q ooo o o ooo oo.o oo 00. 00I o co. o co o.IIIo ooou o oo . ooo00o . cocoa-O Occc..Co domoouuoI Iooooowuo cocoon DIImfln 0 0WM000.0 0w0000 u 0n0000 0I-Mn.00m.0 MM0I. 000." M00WW0I0 00mww0.” 00mwu0. “IIM0M 0000000 000000.0I0000 0. 0 .00II DUO coco 00° COLOQOoO ndd OOOOOOQO OOLOCOQOI NH“ 00: CO. CO. H" II 0 O LODOOO a ' oooo. c an 0.0 000 oomII 8() I'll“. \l'llay It hisiilill Al”..-lll. |I'l' Il‘v'l'll'l 1 Q" N u\ L- k!\ 3") am V1 '4 A. .‘.,. c 40‘. '{flv‘fiflfi4r0 '0‘! oaoomwrcmonmmuuomvom I. C L.) I O . (N 7‘! f'.) ( n (2‘ C) C.) C o C: 7.0 (_ g. 3- L|C (. (J (‘11(‘ f I. CCCC‘ (,‘£» LULUL (‘ C C. firtCLY)(‘VLJ P' InflOOO(30%'|(JJ\~1‘\IL{\\T CT 5. 'tf. '- , a" C t. let7 (‘7 c. C” L.) C. I". C .C..‘..... . C K C. ( I: CC I; 0. CO ICCUGL,‘ II nococrUc canon-«moment o a v Hr‘v-OH C m 0 goo on; (.6 man (.0. uni... (.u but, L.. (U; ..U‘ UUU (.o UUU roO GQCC UJU UUU r.o mg; U.U UOU. ..0 can .c .0. ..u UUU ...cxuo.. g.c owe caoc. capo t 5 $0.0 It'nt.‘ «j coo L)L'-O I.) (_I '_J O (J ( " f" C:- CFC. (2c 4 (DCWC‘QCU'LZ‘C) (300C000 \ Q ocfiounucgomc.c~u\- LA C' fi. g; s. (V (41 ( ,..-‘ 00(‘S‘.( UUUUUUU UUUUUUU UUU UUUU u UU.Un NUUc cc. nU.c pans. uuw UUuU UUUU UU.U .Uaox v4 Ua.q‘ ocuu UU.U mm. .Umox «A. :9": zanmx :04 :uU: mUUUmzmc Um. ow. on. I‘ll. :.;o=-lse. -ae:I;:-|e:I.t,11|ucztz;:é=-ri;;;_ 1. av. on. DU. Ililllla . u mmUUUJUUmdz Uzuovmnmcz.U >U:m2 U UUUU mUUUUU «ma UUUU. . UUUUmMUIJUUUU zwzwmimmMmla UZUUmUzmzzm>o csz Uzmx mUJUUU Uz< .mp4:o< mmzxam m mUxU onuUUU m4a<4U<>U «Um» wa2 «cu mUJDUU czuxawHAUUUnmc «ZUmUmHUUUUUUU zFUz awkm UUU-1UUUUUU.UIAUUUUU._U. q: UUUUUUwUli U3 UU3 U UUUUUU mi:cUUU 3 U I.UUU _U :UUU UU. U UUU UU.U;-UUU11 UUU UUUUUU.U UUUUU3 U UUUUUU.U UUUU3 UUUUU3 U UUU U U U:UUU.U UUU UU.U UUU UUU UUUUUU.U UUUUUU.U UUUUUU.U UUUUU3 UUUUU.U UUUU U.U MUM U. UUUUUU.U UUU UU.U UUU-u UUU UUUUUU.U- UU UU3 U UUUUUU.U-1UUUUU3 lUmUUUU.U UUU3 U.U;4UUU U :UUU U -UUU U.U UUUU. UUU UUUUUU.U UU U.3 U UUZ UU. U UUUUU3 UUUUUU.U UUU ...U UU,_ 3 UUUUUU. UUU U.U UOU UUU UUUUUU.U UU.UUU U UUUUUU.U UUUUU3 UU.U UUU U.U .UUU U. UUUUUU. UUU U.UaIUUm i U.U UUUULU.U,.UUUUU3 U.'UUUUUUuU:sUUUUUU UU UUU.U-;UUU UU.U11UUU UU --UUUUUU. UU U.U UU. U.U UUUUUU.U UU_UUU. U UUcUUU.U UUUUUU U U.U.U U U. U.U UUU UU UUUUUU. UU. U.U UUU ,U UUUUUU.U UU,UUU.U UUUUUU.U UUUUUU U UUU. U UUUUUU.U UUU UU UUUUUU. UU U.U UUU ”U UUUUUU.U-1UU: .UU,U.-UUUUUU.U:;UUU nU UUU U! UUUUUU.U UUU UU UUUUUU IUUUUUU.qleUU1u UU UUUUUU.U UUU U U. U UUUUUU.U UUUUUU.U U UUU. U UUUUU‘.U UUU UU UUUUUU. UUUUUU.U UUU UUU UUUUUU.U UU U U. U UCUUUU.U UUUUUU.U UU. U UUUUUU.U UUU UU UUU UU. .U UU.§ .U.U UUUUUU.U; UUUUUU.U- (UUUUU.U UUUUUU.U:; HUU.Ui UU UU . U‘ UUU UU IUUUUUU U .Ul mU.§ U.U UU.UUU.U UU_UUU. U UUUUUU.U UUUUUU.U UUUUUU.U UUUUUU. U -U UUUUUU. U U UUU U U UUUUUU.U UUU U_U.U UUUUUUuU UUUUUU.U UUUUUU.U UUUU.U.U UUUUUU.U .U UUU U U-,UUUUUU.U23UUUUUU.U.1UUUUUU.anoUUUUUUUxxUUUUanq UUUUUU- ;I UUUUUUmqllaUUUUdaqnidocUUo qdl.i .c UUUUUU.° ooUoUoUo occooo.o oooooo.o ocooco.o cacao“. a oooooo.a as“ U UUUUUU.U:uUmUpUpéniamUUUUU U UUUUUU.U UUUUUU.U UUUUUU.U UUUUUU.U UWMI. U UUUUUQ U ‘UUUUUU.U UUUUU U. U UUUUUU.U UUUUUU.U UUUUUU.U UUUUUU.U UUU lHIllnfllflijlflmflh[lelljlfllflNllllMlLlel‘lH