‘13 Michigan S no University This is to certify that the thesis entitled STAND PARAMETERS AFFECTING MORTALITY FACTORS OF SPRUCE BUDWORM EGGS AND DISPERSING LARVAE presented by WILLIAM PAUL KEMP has been accepted towards fulfillment of the requirements for MLSL degree in ENTOMQLOQY Major professor Date a/iflg 0-7 639 STAND PARAMETERS AFFECTING MORTALITY FACTORS OF SPRUCE BUDWORM EGGS AND DISPERSING LARVAE BY WILLIAM P. KEMP A THESIS sabmitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Entomology 1978 ABSTRACT STAND PARAMETERS AFFECTING MORTALITY FACTORS OF SPRUCE BUDWORM EGGS AND DISPERSING LARVAE BY WILLIAM P. KEMP A study was conducted in Maine to examine the effects of stand characters on population fluctuations of egg and dispersing first- and second-instar spruce budworm (Choristoneura fumiferana (Clemens)). Data included intensive information on stand density, composition, crown closure, crown surface area, basal area, height, and associated vegetation across a 60 plot continuum. Results of data analyses and computer model- ling indicated that open area and non-host tree densities favorably affected budworm parasites Trichogramma minutum Riley, Apanteles fumiferanae Viereck, and Glypta fumiferanae (Viereck). Open areas and non-host tree densities were also found to significantly increase mortality of dispersing budworm in the fall and spring dispersal periods. ACKNOWLEDGEMENTS I would like to extend my sincere thanks to the following people for their contributions to my thesis: To my advisor, Dr. Gary Simmons, whom I will always remember for his warmth, understanding, encouragement, professionalism, and friendship. To my committee, Drs. Fred Stehr, Richard Merritt, Randall Heiligmann and Louis Wilson, all of whom provided valuable ideas and suggestions. To Cheryl, always to Cheryl. ii TABLE OF CONTENTS PAGE LI ST OF TABLES ........ O O O O O O C O O O O O O O O I O O O O ......... O O O ........... Vi LIST OF FIGURES ............... . .................................. viii CHAPTER 1 - CURRENT STATUS OF SILVICULTURAL CONTROL OF SPRUCE BUDWORMCO.....0.00.0.........OCOCOOOOOOOOOO 0000000000 INTRODUCTION O.......OOOOOOOOOO......OOOOOOOOOO0.00.00.00.00. 1 LIFE HISTORY OF THE SPRUCE BUDWORM ............... ....... .... 2 SUSCEPTIBILITY AND VULNERABILITY OF THE FOREST . ............. 3 DISPERSAL LOSS OF BUDWORM LARVAE ............................ 4 CURRENT PHILOSOPHY OF SILVICULTURAL CONTROL ........... ...... 7 SIGNIFICANCE OF THIS STUDY ............ . ................ ..... 7 OBJECTIVES OF THIS STUDY ... ..... ............ ...... .... ...... 8 REFERENCES CITED ................ ...... . ...... ..... .......... 10 CHAPTER 2 - THE INFLUENCE OF STAND FACTORS ON THE PARASITISM OF SPRUCE BUDWORM EGGS BY Trichogramma minutum ....... ESTRACT C O O O O O O O O O O 0 O O O O O 0 O ..... O O O O O O O I O ....... O 00000000000 18 INTRODUCTI ON 0 I O O O O O O O O O O O O O O O O O O I O O O O I O O OOOOOOOOOOOOOOOOOOOO 19 METHODS ......OOOOOOOOOOO0.0.0.............OOOOOOOOOOOOOOO... 21 RESULTS AND DISCUSSION 0.00.00.00.00.......OOOOOOOOOOOOOOOOOO 25 Comparison of High and Mid Crown Densities . ............ 25 Regression Analysis ...................... ............. . 26 iii PAGE Apparent Searching ability of T. minutum .............. 26 Relationships with Forest Succession ......... . ........ 29 ACKNOWLEDGEMENTS ....... .................................... 45 REFERENCES CITED .. ..... . ....... . ........................... 46 CHAPTER 3 - THE INFLUENCE OF STAND FACTORS ON POPULATION LEVELS AND DISPERSAL LOSSES OF INSTAR-I AND INSTAR-II SPRUCE BUDWORM .. ........... . ............. . ABSTRACT 0.0.0..........OOOOOOOOOOOOOO............OOOOOOOOO. 51 INTRODUCTION ..................................... ...... .... 52 METHODS .................................... ................ 54 RESULTS AND DISCUSSION . .................................... 59 Comparison of High and Mid Crown Densities ............ 59 Influences of Stand Factors . .......................... 63 Forest Management Implications ...... . .............. ... 69 ACKNOWLEDGEMENTS ....................... . ...... . .......... .. 84 REFERENCES CITED .......... ....... . ............ . ............ 85 CHAPTER 4 - A MODEL FOR EXAMINING THE EFFECTS OF STAND FACTORS ON SPRUCE BUDWORM LARVAL DISPERSAL ..... ..... ABSTRACT .................... ........................ . ...... 91 INTRODUCTION ....... .......... ..... ... .............. ... ..... 92 METHODS .......................................... .......... 93 Model Description - General ........................... 93 Model Description - Specific ....... ............ . ..... . 94 Analysis ....................... ............ ...... ..... 99 RESULTS AND DISCUSSION ......................... ....... ..... 100 iv PAGE Implications in Forest Management ..................... 102 ACKNOWLEDGEMENTS ...... ..................................... 130 REFERENCES CITED .. ......... . ............................. .. 131 SUMMARY AND RECOMMENDATIONS ...... . .............................. 134 TABLE LIST OF TABLES CHAPTER 2 PAGE 1. T-values and probabilities of spruce budworm viable eggs (instar-I), overwintering larvae (instar-II), and established spring larvae (instar-III) in com- paring high and mid crown balsam fir samples (9.3 m foliage). Island Falls, Vanceboro, Maine, 1976 ....... 34 T-values and probabilities of spruce budworm survival rates during the fall (instar-I) dispersal and the spring (instar-II) dispersal in comparing high and mid crown balsam fir samples. Island Falls, Vanceboro, Maine, 1976 .... ........................... . ........... 36 Variables included in the final mid crown regression equation using survival rates of spring dispersing instar-II budworm (y) and stand parameters (x). Island Falls, Vanceboro, Maine, 1976 .................. 38 Variables included in the final high crown regression equation using survival rates of spring dispersing instar-II budworm (y) and stand parameters (x). Island Falls, Vanceboro, Maine, 1976 .................. 4o Variables included in the final mid crown regression equation using survival rates of fall dispersing instar-I budworm (y) and stand parameters (x). Island Falls, Vanceboro, Maine, 1976 ..... .................... 42 Variables included in the final high crown regression equation using survival rates of fall dispersing instar-I budworm (y) and stand parameters (x). Island Falls, Vanceboro, Maine, 1976 .................. 44 vi TABLE CHAPTER 3 PAGE 1. T-values and probabilities of spruce budworm egg masses, viable eggs, parasitized eggs, non-viable eggs and percent parasitism in comparing high and mid crown balsam fir samples (9.3 m2 foliage). Island Falls, Vanceboro, Maine, 1976 .................. 73 2. Variables included in the final mid crown regression equation using densities of parasitized spruce budworm eggs and stand parameters. Island Falls, Vanceboro, Maine, 1976 .... ............................ 75 3. Variables included in the final mid crown regression equation using an arcsin transformation of percent parasitism and stand parameters. Island Falls, Vanceboro, Maine, 1976 ... .......... . ...... . ........... 77 4. Variables included in the final high crown regression equation using densities of parasitized spruce budworm eggs and stand parameters. Island Falls, Vanceboro, Maine, 1976 ......................... ...... . 79 5. Variables included in the final high crown regression equation using an arcsin transformation of percent parasitism of spruce budworm eggs and stand para- meters. Island Falls, Vanceboro, Maine, 1976 ......... 81 6. Table of variables and the number of regressions in which they were included in the final set of x variables ............ . .............................. 83 vii FIGURE LIST OF FIGURES Program Driver; driving routine for spruce budworm dispersal model ............. ...... ..................... Fall dispersal subroutine for spruce budworm dispersal model ........................................ Spring dispersal subroutine for spruce budworm dispersal model ............... ....... .... .............. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. (Weather = Optimal, Open Area = 10%, Dispersal = 13, Random Spacing) ... ................ . ........ . .................. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Host and non-host aggregated. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open area = 10%, Dispersal = 13, Clumped Spacing) ....... ................ Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Dispersal time decreased. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 10%, Dispersal = 8, Random Spacing) ......................... Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Figure 4 superimposed in hatched lines. (Weather = 75% Cloudy, Open Area = 10%, Dispersal = 13, Random Spacing) .............................. .......... Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Weather during dispersal contained 75% cloudy days. Figure 4 superimposed in hatched lines. (Weather = 75% Cloudy, Open Area = 10%, Dispersal = 13, Random Spacing) .................. ... .......................... viii PAGE 105 107 109 111 113 115 117 119 FIGURE 10. 11. 12. 13. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Open area increased to 20%. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 20%, Dispersal = 13, Random Spacing) ........................ Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Open area increased to 30%. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 30%, Dispersal = 13, Random Spacing) ........................ Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Open area reduced to 0%. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 0%, Dispersal = 13, Random Spacing) .... .................... Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Species proportions were altered to increase black spruce. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 10%, Dispersal = 13. Random Spacing) ........................................ Graph of the effects of non-host percentages of spring and fall budworm. Species proportions of the host component were altered to increase red spruce. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 10%, Dispersal = 13, Random Spacing) ............................................... ix PAGE FIGURE 10. ll. 12. 13. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Open area increased to 20%. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 20%, Dispersal = 13, Random Spacing) . .......... . ............ Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Open area increased to 30%. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 30%, Dispersal = 13, Random Spacing) ... ..... . ............... Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Open area reduced to 0%. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 0%, Dispersal = 13, Random Spacing) ........................ Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Species proportions were altered to increase black spruce. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 10%, Dispersal = 13, Random Spacing) ............................ ............ Graph of the effects of non-host percentages of spring and fall budworm. Species proportions of the host component were altered to increase red spruce. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 10%, Dispersal = 13, Random Spacing) ..... .......................................... ix PAGE CHAPTER 1 CURRENT STATUS OF SILVICULTURAL CONTROL OF THE SPRUCE BUDWORM INTRODUCTION The eastern spruce budworm, Choristoneura fumiferana (Clemens) (Lepidoptera: Tortricidae) is native to North America and is distributed throughout the northern boreal forest from Alberta to Maine. The pre- ferred hosts are balsam fir (Abies balsamea (L.) Mill.) and white spruce (Picea glauca (Muench) Voss) (Greenbank 1963b). Both tree species are harvested for pulpwood and serve as basis for economies in Quebec, New Brunswick and Maine. Therefore, the budworm is competing for a principle resource with man. As a result, it has been a subject of intensive study. Outbreaks of the budworm occur in 30-40 year intervals. The out- breaks themselves may extend only 8-10 years, but the effects are longer lasting (Blais 1965, Fye and Thomas 1963, Ghent et a1 1957, Ghent 1958, 1963, Hatcher 1961, 1963, 1964, Heimberger 1945, McLintock 1955, Williams 1966). In the past, large scale spraying programs controlled the budworm populations. DDT, the principle chemical employed, was sprayed until the early 1960's when, for environmental reasons, it‘s use was restricted. In subsequent years (1968-1977), less residual, more environmentally palatable chemicals were used. These second generation chemicals included organOphosphatesanuicarbamates. Due to reduced dosages of these chemicals, control of budworm populations was no longer attempted. Spraying in the middle 1970's was aimed at keeping mature trees alive until they could be harvested without regard to budworm population control. Throughout this time, concerns were voiced about the effectiveness and environmental 1 impact of the spray programs (Blais 1974). Studies on spruce budworm dynamics and forest character were suggested and in some cases attempted. These new studies were concerned with the interactions of budworm and the forest emphasizing possible unique means of control. LIFE HISTORY OF THE SPRUCE BUDWORM The female spruce budworm moths lay eggs in late July and early August. Oviposition occurs on balsam fir, red spruce (Picea rubens Sarg.), and black spruce (Picea mariana (Mi11.(B.s.P.)),though white spruce is preferred for oviposition (Blais 1957, Wilson 1963). Eggs are deposited about 20 eggs per mass. Normally, there are 200 eggs laid per female. The first-instar, active in the fall, does not feed, but finds an over- wintering site under bark scales, lichens, and in staminate flower cups. The larvae contact overwintering sites through either active or passive dispersal. Dispersal involves the larvae spinning silken threads from branches. The threads are broken and the larvae are carried by atmospheric turbulence to other parts of the forest. After overwintering sites are located, the larvae spin hibernaculae, molt to the second-instar, and overwinter in dispause. The second-instars emerge in the spring, depend- ing on temperature and light. At this time, in most cases, budworm undergo a second dispersal. It is identical to fall dispersal. This results in redistribution of larvae over the entire forest. After dis- persal, the larvae move to branch tips where the current buds are beginning to expand. Here, feeding begins and continues through 6 instars until mid July. Pupation and adult emergence follow shortly. SUSCEPTIBILITY AND VULNERABILITY OF THE FOREST The character of the forest has long been recognized as a fundamental component in the development of spruce budworm outbreaks (Ashley and Stark 1976, Bakuzis and Hansen 1965, Balch 1946, Baskerville 1975, Batzer 1976, Blais 1968, 1973, Brown et a1 1976, Craighead 1923, Graham 1956, Graham and Knight 1965, Morris 1963, Morris and Bishop 1951, Mott 1963, Prebble and Morris 1951, Swaine et a1 1924, Tothill 1923, van Raalte 1972). Foresters recognize two relationships apparent in the forest/budworm in- teraction: (a) the effect of the forest on the budworm and (b) the effect of the budworm on the forest (Mott 1963). Two terms describing these interactions were coined--susceptibility and vulnerability. Susceptibility is the probability of a forested area being attacked by spruce budworm; and vulnerability is the probability of damage resulting from attack (Mott 1963). High susceptibility is directly related to an accumulation of large, contiguous areas of balsam fir. Early cutting recommendations advocated removal of these large accumulations to reduce susceptibility (McLintock 1947, 1954, McLintock and Westveld 1946, Westveld 1946). Other workers developed criteria for identifying highly susceptible areas that might serve as budworm population foci (Balch 1946, Morris and Bishop 1951). Addressing the concept of vulnerability by studying damaged stands, many developed management plans and risk-rating systems for cutting practices (Balch 1946, Batzer 1969, Bean and Batzer 1956, Blais 1964, Frank and Bjorkblom 1973, Graham and Orr 1940, Graham, 1951, 1956, Hatcher 1960, McLintock 1948, 1949, Morris 1958, Morris and Bishop 1951, Swaine et a1 1924, Turner 1952, Westveld 1946, 1954). In none of these instances, however, were budworm population densities and fluctuations related to stand characteristics; only damage was examined. Studies of budworm population dynamics have suggested that stand density and composition influences larval survival (Batzer 1975, Mott 1963). High stand density likely is conducive to survival of dispersing first- and second-instars, while open stands promote large larval survival due to increased crown exposure (Batzer 1976). Mott (1963) postulated that maximum larval survival occurs where the product of large and small larval survival is maximum brought about by mature fir stands having well developed crowns without a significant component of non-host trees or unstocked openings. Studies leading to these postulates, however, did not include actual measurements of dispersal losses immediately following batch or spring emergence from overwintering hibernaculae. DISPERSAL LOSS OF BUDWORM LARVAE Dispersal loss is defined by Miller (Miller 1958) as mortality occurring at two periods: (a) first-instar dispersal in the fall and (b) second-instar dispersal in the following spring. It is consistently high, ranging from 60 - 80% (Miller 1975, Yuill 1958). Probable mortality factors operating during fall dispersal are: (a) air dispersal to non— host material, (b) predation, (c) failure to spin hibernaculae and (d) diapause-free development (Miller 1958). Spring dispersal loss includes: (a) air dispersal to non-host material, (b) predation and (c) failure to establish a feeding site (Miller 1958). Wastage of part of the population on non-host material is considered the major contributor to dispersal loss. A number of subtle stand characteristics can influence mortality factors operating during the period of initial dispersal. At the time of oviposition gravid females are attracted to the tallest, dominant trees due to an apparent positive phototaxis (Mott 1963, Wellington 1948, 1965, Wilson 1963, 1964a, 1964b, Wilson and Bean 1963). The choice of tallest trees appears to have adaptive significance in increasing probability of larvae finding a suitable overwintering site at the time of instar-I dis- persal. Stand microclimate can also influence dispersal mortality. Balsam fir bark is characteristically smooth when trees are young and becomes fissured and flaky with maturity. Physiological maturity also results in production of staminate flowers (Blais 1959, Greenbank 1963a), which leave tiny cup-like bracts upon disintegration. Our observations indicate that extensive lichen mats develOp on the trunk and needle-free portions of branches as balsam fir matures. Bark fissures, staminate flower bracts and lichen mats are common microhabitats where spruce budworm larvae spin overwintering hibernaculae (Miller 1958). In addition, buds which have been excavated by previous generations of budworm are also selected as overwintering sites (Batzer 1968). The microhabitat provided by mature balsam fir trees is less than ideal for many natural enemies of small larvae including parasites, predators and disease (Simmons et a1 1975, VanDenberg and Soper 1975). Crown exposure influences radiation intensity and thus favorably affects body temperature of budworm at times of dispersal (Shepherd 1958, 1959). Tree height appears, thus, to be a potential management variable that bears directly on the population dynamics of the spruce budworm. Indeed, it has been shown that taller stands sustain more defoliation than do shorter stands of comparable age (Mott 1963). Although spruce budworm population dynamics have been extensively studied, only a few facts are known concerning springtime dispersal of second-instars (Shaw and Little 1973). The date of emergence is governed by the accumulation of degree day heat units. Non-parasitized individuals emerge first followed by parasitized larvae (Miller et a1 1971, Shaw and Little 1973). Emergents exhibit positive phototaxis by crawling to branch tips where turbulent winds cause dispersal before they establish feeding sites (Batzer 1968, Henson 1950, Wellington and Henson 1947); non-para- sitized individuals are preferentially dispersed. The velocity of fall is a function of the length of the silk thread spun by the larva and the wind speed (Batzer 1968, Henson 1950). The median dispersal distance from single residual overstory trees is approximately 30 ft (Batzer 1968), although spring dispersal is highly variable with none occurring in some years and dispersal of "several miles" occurring in other years. Indeed, Miller's (1958) data show less variability in fall dispersal than in spring dispersal. This suggests inconsistent spring weather conditions may contribute to variability in spring dispersal. Others have also suggested the same hypothesis (Mott 1963). The specific influence of tree species density and composition on dispersal mortality has not been studied in the same detail that it has been for parasitism (Simmons et a1 1975) . Several authors speculate dispersal to non-host trees likely will result in increased mortality (Jaynes and Speers 1949, Miller 1958, Mott 1963). Yet gross measurements in a variety of stands provided inconclusive data (Miller 1958). Indirect studies examining damage suggests that mixed stands may result in higher dispersal losses (Batzer 1969, Turner 1952). It is established that bud- worm forced to feed on non-preferred hosts develop slower, exhibit more instars and are less fecund than those feeding on foliage of balsam fir (Jaynes and Spears 1949). This infers specific reasons why mortality on non-preferred hosts may be greater. CURRENT PHILOSOPHY OF SILVICULTURAL CONTROL Several silviculture specialists have recently taken the philosophic- al position that the spruce budworm cannot be controlled by silviculture alone (Baskerville 1960, 1975a, 1975b, Batzer 1968, 1976, Gibbs and Blum 1975). Their reasoning suggests that balsam fir tends to regenerate prolifically regardless of the cutting policies employed. Such heavy regeneration of fir is always sufficient to provide ample food for the budworm. Coupled with the mobility of the budworm through long distance moth flights, silvicultural control of susceptibility is unlikely (Baskerville 1975a). Optimism for silvicultural manipulation of vulner- ability has, however, been expressed by these same individuals (Baskerville 1960, 1975a, 1975b, Batzer 1967, 1976, Gibbs and Blum 1975). SIGNIFICANCE OF THIS STUDY The problem is highly complex having resulted from perhaps eons of coevolution of the northern boreal forest and its manager, the spruce budworm. Together the forest and the budworm have developed strategies which maximize the probability of survival of each through time. In our efforts to improve upon this set of complex interactions we have perhaps failed to appreciate the ramifications of competing with the budworm for the rights to manage its domain. In our naivete we have historically relied on simplistic solutions to apparent short-run phenomena only to discover that the budworm/forest ecosystem, through build-in homeostasis, has absorbed our pertubations and continued. In short, simplistic solu- tions were not ample. Silvicultural practices likely cannot and should not be viewed as the ultimate solution to the spruce budworm problem. They, along with past geologic history, climate, weather, parasites, predators, disease, and economics bear directly on the dynamics of the budworm/forest and how effectively we may deal with them--some variables are manageable and others are not! Further, some are entirely manageable and others only partially or indirectly so. Silvicultural practices are among several different controls that, when integrated, may lead us closer to competing well with Choristoneura fumiferana. Obviously when we observe areas sustaining only light de- foliation in the midst of an intense outbreak, there are microdynamics occurring at those locations which we do not fully understand. Past studies contributed to the knowledge of the direct and indirect influence of stand composition and density on the dispersal mortality of newly hatched fall budworm larvae and newly emerged spring budworm larvae-- processes which usually result in 60—80% budworm deaths. OBJECTIVES OF THIS STUDY The primary objective of this study was to examine the effects of stand characteristics on population fluctuations of egg and dispersing first- and second-instar budworms. In order to achieve this goal, the study was separated into 4 parts. A literature review is presented in chapter 1. Chapter 2 contains a study of the effects of stand factors on the egg parasite Trichogramma minutum Riley and the resulting impact on parasitism rates of spruce budworm eggs. In chapter 3, the effects of stand characteristics on population levels and survival rates of fall (instar—I) and spring (instar-II) budworm are examined. Lastly, chapter 4 describes a computerized dispersal model incorporating information from the previous 3 chapters and from the most recent computer modelling techniques. This model was used to examine the effects of a variety of stand and environmental factors on dispersal losses in the fall and the spring. REFERENCES CITED Ashley, M. D., and D. Stark. 1976. Photo field guide for on-the-ground evaluation of spruce budworm damage (Choristoneura fumiferana) on balsam fir (Abies balsamea Mill.) Univ. Me. Sch. For. Res. and State Dep. Cons. Bur. Forestry. 20 pp. Bakuzis, E. V, and H. L. Hansen. 1965. Balsam Fir: Abies balsamea (L.) Mill. A monographic review. Univ. Minn. Press, Minneapolis. 445 pp. Balch, R. E. 1946. The spruce budworm and forest management in the Maritime Provinces. Can. Dep. Agr., Entomol. Div. Processed Pub. 60. 7 pp. Baskerville, G. L. 1960. Mortality in immature balsam fir following severe budworm defoliation. Forest Chron. 36:342-345. Baskerville, G. L. 1975a. Spruce budworm: super silviculturist. Forest Chron. 51:4-6. Baskerville, G. L. 1975b. Spruce budworm: the answer is forest manage- ment: or is it? Forest Chron. 51:23-26. Batzer, H. O. 1967. Spruce budworm defoliation is reduced most by commercial clearcutting. USDA Forest Serv. Res. Note NC-36, N. Cent. Forest Exp. Sta., St. Paul, Minn. 4 pp. Batzer, H. O. 1968. Hibernation site and dispersal of spruce budworm larvae as related to damage of sapling balsam fir. J. Econ. Entomol. 61:216-220. 10 ll Batzer, H. O. 1969. Forest character and vulnerability of balsam fir to spruce budworm in Minnesota. Forest Sci. 15:17-25. Batzer, H. O. 1976. Silvicultural control techniques for the spruce budworm. In Chansler, J.F., and W. H. Klein, pgs. 110-116, Pro- ceedings of a symposium on the spruce budworm, November 11-14, 1974, Alexandria, Virginia. USDA Forest Serv. Misc. Pub. No. 1327. 188 pp. Bean, J. L., and H. O. Batzer. 1956. A spurce budworm risk rating for the spruce-fir types in the Lake States. USDA For. Serv. Tech. Note 453, Lake States Froest Exp. Sta. St. Paul, Minn. 2 pp. Blais, J. R. 1957. Some relationshipscflfthe spruce budworm to black spruce. Forest Chron. 33:364-372. Blais, J. R. 1959. The vulnerability of balsam fir to spruce budworm attack in northwestern Ontario, with special reference to the physiological age of the tree. Forest Chron. 34:405-422. Blais, J. R. 1964. Account of a recent spruce budworm outbreak in the Laurentide Park Region of Quebec and measures for reducing damage in future outbreaks. Forest Chron. 40:313-323. Blais, J. R. 1965. Spruce budworm outbreaks in the past three centuries in the Laurentide Park, Quebec. Forest Sci. 11:130-138. Blais, J. R. 1968. Regional variation in susceptibility of eastern North American forests to budworm attack based on history of out- breaks. Forest Chron. 44:17-23. Blais, J. R. 1973. Control of spruce budworm: current and future strategies. Entomol. Soc. Am. Bull. 19:208-213. Blais, J. R. 1974. The policy of keeping trees alive via spray operations may hasten the recurrence of spruce budworm outbreaks. Forest Chron. 50:19-21. 11 Batzer, H. O. 1969. Forest character and vulnerability of balsam fir to spruce budworm in Minnesota. Forest Sci. 15:17-25. Batzer, H. O. 1976. Silvicultural control techniques for the spruce budworm. In_Chansler, J.F., and W. H. Klein, pgs. 110-116, Pro- ceedings of a symposium on the spruce budworm, November 11-14, 1974, Alexandria, Virginia. USDA Forest Serv. Misc. Pub. No. 1327. 188 pp. Bean, J. L., and H. O. Batzer. 1956. A spurce budworm risk rating for the spruce-fir types in the Lake States. USDA For. Serv. Tech. Note 453, Lake States Froest Exp. Sta. St. Paul, Minn. 2 pp. Blais, J. R. 1957. Some relationshipscufthe spruce budworm to black spruce. Forest Chron. 33:364-372. Blais, J. R. 1959. The vulnerability of balsam fir to spruce budworm attack in northwestern Ontario, with special reference to the physiological age of the tree. Forest Chron. 34:405-422. Blais, J. R. 1964. Account of a recent spruce budworm outbreak in the Laurentide Park Region of Quebec and measures for reducing damage in future outbreaks. Forest Chron. 40:313-323. Blais, J. R. 1965. Spruce budworm outbreaks in the past three centuries in the Laurentide Park, Quebec. Forest Sci. 11:130-138. Blais, J. R. 1968. Regional variation in susceptibility of eastern North American forests to budworm attack based on history of out- breaks. Forest Chron. 44:17-23. Blais, J. R. 1973. Control of spruce budworm: current and future strategies. Entomol. Soc. Am. Bull. 19:208-213. Blais, J. R. 1974. The policy of keeping trees alive via spray Operations may hasten the recurrence of spruce budworm outbreaks. Forest Chron. 50:19-21. 12 Brown, M. W., F. B. Knight and J. B. Dimond. 1976. Stand composition and susceptibility to spruce budworm epidemics. Univ. Me. Sch. Forest Res. Tech Note No. 61. 4 pp. Craighead, F. C. 1923. A brief summary of the budworm investigations in Canada. J. Forestry 21:134-138. Frank, R. M., and J. C. Bjorkblom. 1973. Silvicultural guide for spruce- fir in the Northeast. USDA Forest Serv. Gen. Tech. Rep. NE-6 North- east. Forest Exp. Sta., Upper Darby, Pa. 29 pp. Fye, R. E., and J. R. Thomas. 1963. Regeneration of balsam fir and spruce about fifteen years following release by spruce budworm attacks. Forest Chron. 39:385-397. Ghent, A. W., D. A. Fraser and J. B. Thomas. 1957. Studies of regene- ration in forest stands devastated by the spruce budworm. I. Forest Sci. 3:184-207. Ghent, A. W. 1958. Studies of regeneration in forest stands devastated by the spruce budworm. II. Forest Sci. 4:135-146. Ghent, A. W. 1963. Studies of regeneration in forest stands devastated by the spruce budworm. III. Forest Sci. 9:295-310. Gibbs, C. B., and B. M. Blum. 1975. Silviculture and the spruce budworm in Maine. Me. Forest Rev. 8:13-15. Graham, S. A., and L. W. Orr. 1940. The spruce budworm in Minnesota. Univ. Minn. Agr. Exp. Sta., Tech. Bull. 142. 27 pp. Graham, S. A. 1951. Developing forests resistant to insect injury. Sci. Mon. 73:235-244. Graham, S. A. 1956. Hazard rating of stands containing balsam fir accord- ing to expected injury by spruce budworm. Univ. Mich. Dep. Forest. 13. 2 pp. 13 Graham, S. A., and F. B. Knight. 1965. Principles of Forest Entomology. Ed. 4. McGraw-Hill, New York. 417 pp. Greenbank, D. O. 1963a. Staminate flowers and the spruce budworm. In R. F. Morris (ed.), pgs. 208-218, The Dynamics of epidemic spruce budworm populations. Entomol. Soc. Can. Mem. 31:1-332. Greenbank, D. O. 1963b. Host species and the spruce budworm. In_R. F. Morris (ed.), pgs. 219-223, The dynamics of epidemic spruce budworm populations. Entomol. Soc. Can. Mem. 31:1-332. Hatcher, R. J. 1960. Development of balsam fir following a clearcut in Quebec. Can. Dep. North. Affairs and Nat. Resourc., Forest Resour. Div., Tech. Note 87. 22 pp. Hatcher, R. J. 1961. Partial cutting balsam fir stands on the Epaule River watershed, Quebec. Can. Dep. Forest., Forest Res. Branch Tech. Note 105. 29 pp. Hatcher, R. J. 1963. Effects of birch dieback and spruce budworm on forest development, Forest Section L. 6, Quebec Can. Dep. Forest., Forest Res. Branch, Pub. 1014. 16 pp. Hatcher, R. J. 1964. Spruce budworm damage to balsam fir in immature stands, Quebec. Forest Chron. 40:372-383. Heimberger, C. C. 1945. Comment on the budworm outbreak in Ontario and Quebec. Forest Chron. 21:114-126. Henson, W. R. 1950. The means of dispersal of the spruce budworm. Ph.D. Thesis. Yale Univ., New Haven, Conn. Jaynes, H. A., and C. F. Speers. 1949. Biological and ecological studies of the spruce budworm. J. Econ. Entomol. 42:221-225. McLintock, T. F. 1947. Silvicultural practices for control of spruce budworm. J. Forestry 45:655-658. 14 McLintock, T. F. 1948. Evaluation of tree risk in the spruce-fir region of the Northeast. Iowa State Coll. J. Sci. 22:415-419. McLintock, T. F. 1949. Mapping vulnerability of spruce-fir stands in the Northeast to spruce budworm attack. USDA Forest Serv. Sta Pap. 21, Northeast Forest Exp. Sta., Upper Darby, Pa. 20 pp. McLintock, T. F. 1954. Factors affecting wind damage in selectively cut stands of spruce and fir in Maine and northern New Hampshire. USDA Forest Serv. Sta. Pap. NE-70, Northeast, Forest Exp. Sta., Upper Darby, Pa. 17 pp. McLintock, T. F. 1955. How damage to balsam fir develops after a spruce budworm epidemic. USDA Forest Serv. Sta. Pap. 75, Northeast Forest Exp. Sta., Upper Darby, Pa. 18 pp. McLintock, T. F. and M. Westveld. 1946. Some financial aspects of removal of overmature balsam fir as a budworm control measure. USDA Forest Mgmt. Pap. No. 1, Northeastern Forest Exp. Sta., Upper Darby, Pa. 8 pp. Miller, C. A. 1958. The measurement of spruce budworm populations and mortality during the first and second larval instars. Can. J. Zool. 36:409-422. Miller, C. A., D. C. Eidt and G. A. McDougall. 1971. Predicting spruce budworm development. Dep. Env. Can. Forestry Serv., Bi-monthly Res. Notes 27:33-34. Morris, R. F. 1958. A review of the important insects affecting the spruce-fir forest in the Maritime Provinces. Forest Chron. 34: 159-189. 15 Morris, R. F. 1963. Resume. In_R. F. Morris (ed.), Pgs 311-320, The dynamics of epidemic spruce budworm populations. Entomol. Soc. Can. Mem. 31:1-332. Morris, R. F., and R. L. Bishop. 1951. A method of rapid forest survey for mapping vulnerability to spruce budworm damage. Forest Chron. 27:171-178. Mott, D. G. 1963. The forest and the spruce budworm. I2_R. F. Morris (ed.), pgs. 189-202, The dynamics of epidemic spruce budworm pop- ulations. Entomol. Soc. Can. Mem. 31:1-332. Prebble, M. L., and R. F. Morris. 1951. Forest entomology in relation to silviculture in Canada. Part III. The spruce budworm problem. Forest Chron. 27:14-22. Shaw, J. G., and C. H. A. Little. 1973. Dispersal of second instar spruce budworm. Dep. Env. Can Forestry Serv., Bi-monthly Res. Notes 29:30-31. Shepherd, R. F. 1958. Factors controlling the internal temperature of spruce budworm larvae, Choristoneura fumiferana (C1em.). Can. J. 2001. 36:779-786. Shepherd, R. F. 1959. Phytosociological and environmental characteristics of outbreak areas of the 2-year cycle spruce budworm, Choristoneura fumiferana. Ecology 40:608-620. Simmons, G. A., D. E. Leonard and C. W. Chen. 1975. Influence of tree species density and composition on parasitism of the spruce budworm, Choristoneura.fwmiferana (C1em.). Env. Entomol. 4:832-836. l6 Swaine, J. M., F. C. Craighead, and J. W. Bailey. 1924. Studies on the spruce budworm. Part II. General bionomics and possibilities of prevention and control. Can. Dep. Agr. Bull. 37. 91 pp. Tothill, J. D. 1923. Notes on the outbreaks of spruce budworm, forest tent caterpillar and larch sawfly in New Brunswick. Acadian Entomol. Soc. Proc. 8:172-182. Turner, K. B. 1952. The relation of mortality of balsam fir caused by the spruce budworm to forest composition in the Algoma forest of Ontario. Can. Dep. Agr. Pub. 875. 107 pp. VanDenberg, J. V., and R. S. Soper. 1975. Isolation and identification of Entomopthora spp Fres. (Phycomycetes: Entomophthorales) from the spruce budworm Choristoneura fumiferana Clem. (Lepidoptera: Tortricidae). J. New York Entomol. Soc. 83:254-255. Van Raalte, G. D. 1972. "Do I have a budworm-susceptible forest?" Forest Chron. 48:190-192. Wellington, W. G. 1948. The light reactions of the spruce budworm, Choristoneura fumiferana Clemens (Lepidoptera: Tortricidae). Can. Entomol. 80:56-82. Wellington, W. G. 1965. The use of cloud patterns to outline areas with different climates during population studies. Can. Entomol. 97:617- 631. Wellington, W. G., and W. R. Henson. 1947. Notes on the effects of physical factors on the spruce budworm, Choristoneura fumiferana (Clem). Can. Entomol. 79:168-170, 195. Westveld, M. 1946. Forest management as a means of controlling the spruce budworm. J. Forestry 44:949-953. 17 Westveld, M. 1953. Ecology and silviculture of the spruce-fir forest of eastern North America. J. Forestry 51:422-430. Westveld, M. 1954. A budworm vigor resistance classification for spruce and balsam fir. J. Forestry 52:11-24. Williams, C. B. 1966. Differential effects of the 1944—56 spruce budworm outbreak in eastern Oregon. USDA Forest Serv. Res. Pap. PNW-33, Pac. Northwest Forest and Range Exp. Sta., Portland, Oreg. 16 pp. Wilson, L. F. 1963. Host preference for oviposition by the spruce bud- worm in the Lake States. J. Econ. Entomol. 56:285-288. Wilson, L. F. 1964a. Oviposition site of the spruce budworm, Choristoneura fumiferana, modified by light. Ann. Entomol. Soc. Am. 57:643-645. Wilson, L. F. 1964b. Observations on the geo-orientation of spruce bud- worm, Choristoneura fumiferana, adults. Ann. Entomol. Soc. Am. 57: 645-648. Wilson, L. F., and J. L. Bean. 1963. Site of spruce budworm masses on their preferred hosts in the Lake States. J. Econ. Entomol. 56:574- 578. Yuill, J. S. 1958. A study of the migration habits of first instar spruce budworm larvae by means of radioactive tracers. USDA Forest Serv. Forest Ins. Lab. Spec. Rep. Bv 58-1, mimeo, Beltsville, Md. 5 pp. CHAPTER 2 THE INFLUENCE OF STAND FACTORS ON PARASITISM OF SPRUCE BUDWORM EGGS BY Trichogramma minutum ABSTRACT Data included intensive information on stand density, composition, age, crown closure, crown surface area, basal area, height, and associ- ated vegetation across a 60 plot ecological continuum in Maine. T-test results showed significantly higher numbers of egg masses, viable eggs, non-viable eggs, parasitized eggs, and percent parasitism in the upper crowns when compared to mid-crown levels (P<0.05). Regression analyses indicated that parasitism rates rose with increasing density of non- budworm host tree species, especially hardwoods such as sugar maple (Acer saccharum Marshall), yellow birch (Betula alleghaniensis Britton), paper birch (B, papyrifera Marshall), red maple (A, rubrum L.), hep hornbeam (Ostrya virginiana (Miller) Koch), and big tooth aspen (Populus grandidentata Michaux). 18 INTRODUCTION Studies of spruce budworm population dynamics have suggested that stand density and composition influence larval budworm survival (Morris 1963). Simmons et al (1975) showed an association between spruce budworm parasitism rates and tree density and stand composition. Para- sitism of spruce budworm pupae increased as stand density decreased. The observation held regardless of stand composition. Trichogramma minutum, the only known egg parasite of budworm, is widely distributed in North America and preys on eggs of at least 6 orders of insects (Muesebeck et a1 1951). These hosts are found on a wide variety of plant and tree species. In association with spruce budworm, T, minutum functions as a facultative parasite (McGugan and Blais 1959, Miller 1953, Morris 1963). A detailed account of the biology of T, minutum is given by J. L. Sweetman (1958). The family consists exclusively of egg parasites which attack a wide variety of insect hosts. This species spends the winter as a partially developed larva in the host egg. The size of the host egg determines the number of individuals that may develop. The females mate shortly after hatching but are capable of reproducing partheno- genetically. Parasitism by Trichogramma produces an immediate cessation of the normal development of the host embryo. Sweetman (1958) states that about 40-50 eggs per female are laid. Temperature is an important 19 20 factor in determining the length of the life cycle of the parasite. For individuals, the range is 7 to 75 days. During the summer, development from egg to adult is 9 to 16 days. The rate of change in the length of the life cycle at average temperatures from 15.50 C to 21° C is an increase or decrease of one day for each 0.70 C decrease f or increase in the average temperature. Approximately 13-52 generations i can occur each year (Sweetman 1958). Previous studies of spruce budworm parasitism showed conflicting L results as to the importance of parasites. Miller (1953) stated that the combined action of parasites and other natural factors at low bud- worm levels are effective in keeping the populations down. Dowden and Carolin (1950) reported wide differences in percentages of parasitism in a similar study. Jaynes and Drooz (1952), however, suggested para- sites to be of limited importance in the spruce budworm problem in Maine. This report stresses the significance of stand variables on parasitism of budworm by T, minutum. METHODS This study was conducted in central Aroostook and eastern Penobscot counties of Maine (Leonard and Simmons 1974). Balsam fir (Abies balsamea (L.) Miller) predominated and other softwood species associated were white spruce (Picea glauca (Muenchhausen Voss)), red spruce (P, rubens Sargent), hemlock (Tsuga canadensis (L.) Carr), white pine (Pinus Hardwoods strobus L.) and northern white cedar (Thuja occidentalis L.). <:onsisted of yellow birch (Betula alleghaniensis Britton), white birch (g, papyrifera Marshall), grey birch (B, populifolia Marshall), sugar nmaple (Acer saccharum Marshall), red maple (A, rubrum L.), american elm (tJlmus americana L.), white ash (Fraxinus americana L.), big tooth aspen (]?opulus grandidentata Michaux). american beech (Fagus grandifolia Ialixhart) and hop hornbeam (Ostrya virginiana (Miller) Koch). Sixty sites were chosen for data collection. An attempt was made t:<> sample a continuum of forest types from nearly pure hardwood to pure Softwood. Since balsam fir is the primary host for budworm, all sampled trees were balsam fir. Plots were 2 chains in diameter (40.2 m), with CeP-l'lters assigned to a single, mature balsam fir. To sample a continuum, £>J=C>tzs included 1-4 additional balsam fir trees. These additional (secondary) trees were on bearings north, south, east, and west of the center tree. The continuum sampled ranged from 1 primary balsam fir EB‘JCITthaunded by non-host species to 1 primary balsam fir with 4 21 22 additional secondary balsam fir at the 4 points of the compass. These plots corresponded to predominately hardwood and predominately softwood areas respectively. Such plot selection functioned to provide forest successional stages for study. Earlier, more diverse ecological conditions were represented by plots containing 1 central balsam fir tree surrounded by non-host hardwood tree species. Later, less diverse successional stages were represented by plots containing almost entirely balsam fir. which is considered to be a sub-climax or climax species in the northern forest type of Maine (Eyre 1954). Full branch samples were taken at 2 levels, high crown and mid crown, from both the central and secondary sample tree. Five sample branches were taken from each level, measured for length and width, and transported to the lab for study. In the lab, branches were cut up and examined for egg masses. Cross-tabulation of egg mass length with the numbers of egg rows was used to derive the number of eggs per egg mass (Leonard et a1 1973). Data collected from the sample branches consisted of egg masses, viable eggs, parasitized eggs, non-viable eggs, percent parasitism, totals for each branch, and the total for the 5 branches at each sampling level. Approximately 4,000 egg masses were examined from 1,890 branches containing 57,568 eggs. There was a great deal of variability in branch size. Since the crowns of balsam fir tend to be conical, high crown branches tended to be smaller than mid crown branches. Therefore, all observations were standardized to 9.3 m2 of balsam fir foliage (Simmons 1973). 23 T-test analysis and stepwise multiple regression were used with both the upper and mid crown data to identify possible contributing factors to parasitism differences. Two regressions were run at upper and mid levels for each primary tree (60) using the number of para- sitized eggs and percent parasitism as dependent variables respectively. Independent variates included data from each sample tree such as height, dbh, crown length, crown diameter, crown closure, age, crown volume, and crown area. Other independent variables used were stand related factors such as basal areas of all tree species found in the plot. These basal areas were separated into 2 groups, overstory (dominant or co-dominant) and understory (sub-dominant or repressed) trees. For example, 1 variate was overstory sugar maple and another variate was understory sugar maple. The least squares method of regression analysis was extended to a stepwise, forward inclusion, multiple regression model as described by Nie et a1 (1970) and Dixon and Massey (1969). Regressions were cal- culated using the Statistical Package for the Social Sciences default values of (n, f, t) converging on models of about 30 (x) variables (Nie et a1 1970). Therefore, the inclusion value was set at a maximum of 20 (x) variables. The F value and tolerance limit were commanded to be 1.00 and 0.01 respectively. In an effort to find the important factors influencing densities of T, minutum, the regression models were analyzed collectively. Of all the variables entered into the regression models, analysis was concentrated on those that were included in the final equation of 2 24 or more regressions (Table 6). These represented the variables exerting the most influence on T, minutum populations. Central trees in all plots were used as center points for collecting stand data. As a result, all regression analyses were based on central tree data. Further, only those trees containing budworm egg masses were used in the anlayses. Since T, minutum requires a host, to include stand information from areas where there were no spruce budworm would only serve to obscure results. RESULTS AND DISCUSS ION Comparison of High and Mid Crown Densities The mean number of egg masses, viable eggs, non-viable eggs, and parasitized eggs were significantly higher in the high crown when compared to mid crown samples (P<0.05) (Table 1). Percent parasitism was also significantly higher in the high crown level than at the mid crown level (P<0.05) (Table 1). Thomas (1966) proposed this, but lacked statistical support. Mean percent parasitism for T, minutum was 28.2% for the high crown samples and 19.2% for the mid crown samples and ranged from 0% to 100% in both crown heights. These rates were higher than previous reports. Dowden et al (1948) reported a mean percent parasitism of 12%, with a range of 3% to 23%. Thomas (1966) also noted very low parasitism percentages in a budworm infestation in Maine. Standard deviations of egg densities from high crown samples were higher than the standard deviations from mid crown samples. Since the standard deviation is a measure of dispersion, this indicates that the high crown data tended to be more unstable. This should be taken into account in data analysis. Morris (1955) suggested this relationship and indicated that mid crown samples were best suited for budworm population prediction. Thomas (1966) showed, however, that g, minutum 25 26 concentrated its attacks in the high portion of the crown. The dual sampling thus served as a means of checking for consistency of results. Regression Analysis Two mid crown and 2 high crown regressions were calculated using the density of parasitized eggs and arcsin transformation of percent parasitism as dependent variables respectively (Tables 2,3,4,5). As can be seen, the multiple R values for the mid crown regressions were 0.89 and 0.86, the 2 highest in the study. The overall accuracy of the prediction abilities of the equations is reflected by the re- spective R2 values of 0.79 and 0.74. Because of decreased stability of the high crown samples, multiple R and R2 values tended to be lower. The multiple R values of the 2 high crown regressions were 0.60 and 0.76 respectively. The basal areas of overstory yellow birch, red spruce, sugar maple, big tooth aspen, paper birch, red maple, and basal areas of understory red spruce and hop hornbeam were all positively correlated and included in the final models of at least 2 of the 4 final regressions (Table 6). This indicated the importance of these species in increasing T, minutum densities and percent parasitism rates. Apparent Searching Ability of T. minutum Laing (1937, 1938) noted the frequency with which the parasite found nearby host eggs varied inversely with the distance between eggs. When Trichogramma leaves an egg it has just parasitized or examined, it 27 does so in a twisted, winding motion away from the egg. Such non-random search may be why higher parasitism occurs in the high crown areas. Here, the branches are much smaller than at the mid crown heights, providing less surface area to search. Other studies showed that parasites concentrate search at particular crown levels (Shahjahan and Streams 1973, Dodge 1961). Therefore, the success rate of T, minutum should be increased in high crowns. Indeed this was the case. In recent years, there has been increasing concern about inter- specific chemical communication in insects. Brown et a1 (1970) described these chemicals as kairomones. Several studies have shown the importance of chemical communication in ecological balance and management of insect hosts and their parasites (Jones et al 1973, Hendry et al 1973, Lewis et al 1975a, Lewis et al 1975b). Scales left by the female Heliothis zea (Boddie) at oviposition were the kairomone source eliciting a host-searching response from female T, evanescens Westwood (Lewis et a1 1972). These kairomones, normally emitted by the host, apparently serve many functions. Kairomones may elicit different behaviors including attracting parasites to infestations of host insects, promoting close range searching for hosts, and causing oviposition in hosts (Brown et a1 1970, Whittaker and Feeny 1971). In addition, percent parasitism, progeny per female, and female longevity increased significantly when T, pretiosum Riley females were exposed to H, Eg§_scales (Nordlund et a1 1976). These studies may help explain our findings of increased para- sitism in the high crown. Most authors agree that female budworm 28 moths tend to deposit eggs on foliage of taller spruce and fir trees (Mott 1963, Wellington 1948, Wilson 1963, Wilson 1964a, Wilson 1964b, Wilson and Bean 1963). It follows that increased activity of adult females and higher egg mass densities in high crowns should increase amounts of kairomones present. This may cause T, minutum to spend more time searching in the area of the high crown. As a result of the effects of kairomones on the parasite Trichogramma, total production of progeny may be increased through increased egg deposition and greater longevity. Available light may also be a factor since T, minutum uses sight to a limited extent in searching (Laing 1938, Thomas 1966). Laing (1938) observed that both very bright light and dim light hindered rates of success in host searching by Trichogramma. He noted that the highest percentage of success was exhibited in normal daylight during June and July. Parasitism occurred in this study during the month of July. High crown areas receive the greatest amount of sunlight with mid and low crowns receiving much less light, depending on stand density (Spurr and Barnes 1973). Factors such as this may aid T, minutum in searching high crown areas. Our regression analyses indicated that height is positively cor- related to T, minutum parasitism of budworm eggs (Tables 2,3,4,5). We cannot determine at this time whether T, minutum is attracted to the high crowns by environmental conditions or increased host densities. However, it seems likely that a complex of host-parasite relationships, combined with macro- and micro-environmental factors is operating. 29 Relationships with Forest Succession It is known that in parasite complexes, a very important factor which eliminates many species from the potential host list is the failure of the habitats of the 2 species to coincide (Doutt 1959, 1964). Musebeck et a1 (1951) listed over 100 species (6 orders) of insects parasitized by T, minutum, many of which occur on vegetation and tree species associated with the northern forest of Maine. The regression analyses indicated that parasitism increased as the diversity of tree species increased. It follows, therefore, that by increasing the alternate host habitats, T, minutum will increase. Organization of parasite communities results from the interaction of properties of the colonizing species and the changing properties of the environments they exploit (Price 1973b). Further, it has been suggested that parasite complexes are likely to evolve from those species with high reproductive abilities in the early stages, to those with high competitive abilities as succession proceeds (Force 1972). It is clear that the increasing density of hardwood tree species such as sugar maple, hop hornbeam, red maple, big tooth aspen, yellow birch, and paper birch represent the earlier, more diverse stages of succession in the northern forest. Dense stands of mature balsam fir represent later, less diverse successional stages of the forest. Based on these criteria, T, minutum seems to fit into middle, and to some extent, later successional stages. Regression analyses and prior discussion suggest that T, minutum has a high reproductive potential and is most effective at earlier more diverse stages of forest succession. In later successional stages, where predominantly balsam fir exists, 30 T, minutum generally shows lower numbers (Dowden and Carolin 1950, Jaynes and Drooz 1952). Using Force's description of parasite com- plexes, when compared to other parasites, T, minutum does not seem to exhibit necessary competitive abilities to aid it in later suc- cessional stages. Due to a high reproductive potential, as well as its searching ability, it can compete effectively at pre-climax successional stages. In all regressions, crown closure was the only important variable negatively correlated with parasitized egg density and percent para- sitism. It is known that denser forests do not support ground flowers, shrubs and herbaceous plants (Simmons et al 1975). It has been re- ported that small openings on the forest floor support a wide variety of vegetation and a great diversity of habitat (Leopold 1933). This vegetation is important in supporting a variety of alternate hosts for T, minutum. Further, since the spruce budworm does not overwinter as an egg, T, minutum apparently must find another host egg for over- wintering. This may be one reason that T, minutum does not exhibit high populations in late successional stages of a forest where plant and insect diversities are low. Simmons et a1 (1975) suggested that parasite densities were closely related to forest composition for several other spruce budworm predators and parasites. Price (1973a) cites three major factors in- fluencing the organization of a parasite community: (1) successional trends lead to more favorable conditions for a parasite, but host plant densities may influence host densities; (2) host density may fluctuate 31 yearly, independent of host plant densities, but still exert organizing influences on the parasite complex; (3) there is a reasonable pre- dictable succession of host abundances in each generation, from relatively high in the egg stage to low in the adult stage. As both Simmons et a1 (1975) and Price (1973b) indicate, the major factor influencing the T, minutum parasite complex is forest succession. I At both ends of the successional continuum, pioneer and climax, T, minutum is offered fewer alternte host species due to the limited number of alternate host habitats. As a result, T, minutum must rely on its searching technique and its own powers of reproduction to compete in these extreme situations. Analysis shows that T, minutum population levels can be increased by moving toward the middle of the successional continuum with its increasing canopy openings and by the introduction of other non-climax tree species. By interrupting the natural successional trend of a singular dominant climax tree species (balsam fir) in many areas, we can manage for higher budworm mortality due to T, minutum. It is clear that tree species such as sugar maple, yellow birch, paper birch, red maple, big tooth aspen, red spruce and others (depending on local site conditions) are important in establishing the necessary plant and insect diver- sities needed to support healthy T, minutum populations. Managing the forest for increased diversity of plant and tree species would also positively affect other parasite and predator complexes of the spruce 32 budworm (Simmons et a1 1975). It would follow that the combined effects could significantly influence the budworm population in the northern forest type of Maine. 33 Table 1. T-values and probabilities of spruce budworm egg masses, viable eggs, parasitized eggs, non-viable eggs and percent parasitism in comparing high and mid crown balsam fir 2 samples (9.3 m foliage). Island Falls, Vanceboro, Maine, 1976. 34 ooalo vm.mm ma.ma mma Emfluflmmumm unwoumm :3ouo UHE ooo.ovm .ohm mm.~ ooano mm.om vm.mm ems smauhmmnmm unmouwd czouo soar Halo ma.oa mo.m mma mmmm manmh>ucoc csouo was mo.ova .mhm om.a Haano vo.mH mm.o ama momm magma>ncoc nacho roan omaso oo.mma mo.vm mmH mama cmNhuhmmumd czouo was mooo.ovm .mhm mm.m mamauo mo.msm am.msa has momw conauammumm cacao nah: mmmauo om.0mm mH.smm was momm mabmh> cache the Nooo.ovm .mhm mm.q mammuo va.~oo am.omm sma mama magma> asouo rah: moauo mm.sm mH.mm mma mommwe wow cacao was mooo.ovm .mflm mm.m mavlo mm.hm mo.Hm nma mommma moo csouo saws .noum HHmUIH woam>lfi mosmm m .M z mmsouw 33 Table l. T-values and probabilities of spruce budworm egg masses, viable eggs, parasitized eggs, non-viable eggs and percent parasitism in comparing high and mid crown balsam fir 2 samples (9.3 m foliage). Island Falls, Vanceboro, Maine, 1976. 34 ooauo vm.mm os.ma mma smauhmmumm unmouwd nacho was woo.ova .mam mm.m coauo mm.om 4m.mm nma smauhmmumm scanned czouo nah: Halo ma.oa mo.m mma mama manmh>ucoc esouo was mo.ova .mam om.a Hague vo.mH mm.6 nma meow magma>ncoc czouo roan omauo oo.mma mo.vm was meow omuhuhmmuma esono was mooo.ovm .mam mm.m mvmauo mo.msm um.msa sma mama vanguammumm czouo roar mmmauo om.omm ma.amm mma meme manmh> czouu was Nooo.ova .mhm mm.v mammuo vs.mmo «m.0mm smH mama magma> cacao nah: Neale mm.nm mH.m~ mma mmmmme mom cacao was mooo.ovm .mhm mm.m mHVIo mm.hm mo.am hma mounds moo CBOHU sow: .noum Hamula woam>le cocoa m .M z mmsouw 35 Table 2. Variables included in the final mid crown regression equation using densities of parasitized spruce budworm eggs and stand parameters. Island Falls, Vanceboro, Maine, 1976. 36 Variable Multiple R R7- Overstory yellow birch 0.50 0.25 Overstory sugar maple 0.65 0.42 Understory hop hornbeam 0.71 0.51 Crown closure 0.73 0.54 Understory red spruce 0.76 0.57 Overstory hemlock 0.78 0.60 Height 0.80 0.65 Age 0.84 0.70 Dbh 0.85 0.73 Overstory red maple 0.87 0.77 Overstory red spruce 0.88 0.77 Total hardwood basal area 0.89 0.78 Understory balsam fir 0.89 0.79 37 Table 3. Variables included in the final mid crown regression equation using an arcsin transformation of percent parasitism and stand parameters. Island Falls, Vanceboro, Maine, 1976. 38 Variable Multiple R R2 Overstory sugar maple 0.45 0.21 Understory hop hornbeam 0.57 0.32 Overstory yellow birch 0.64 0.41 Height 0.70 0.50 Crown closure 0.75 0.56 Age 0.77 0.60 Total softwood basal area 0.80 0.64 Understory hemlock 0.82 0.67 Crown volume 0.83 0.69 Overstory red maple 0.84 0.71 Overstory big tooth aspen 0.86 0.74 39 Table 4. Variables included in the final high crown regression equation using densities of parasitized spruce budworm eggs and stand parameters. Island Falls, Vanceboro, Maine, 1976. 4O Variable Multiple R R2 Crown closure 0.42 0.17 Overstory yellow birch 0.50 0.25 Understory red spruce 0.55 0.30 Overstory red spruce 0.57 0.34 Overstory paper birch 0.59 0.35 Overstory big tooth aspen 0.60 0.37 41 Table 5. Variables included in the final high crown regression equation using an arcsin transformation of percent parasitism of spruce budworm eggs and stand parameters. Island Falls, Vanceboro, Maine, 1976. 42 Variables Multiple R R2 Overstory sugar maple 0.41 0.17 Overstory paper birch 0.54 0.29 Crown closure 0.63 0.39 Understory hop hornbeam 0.67 0.45 Overstory red spruce 0.68 0.47 Understory sugar maple 0.70 0.49 Understory yellow birch 0.74 0.54 Overstory balsam fir 0.74 0.56 Height 0.76 0.57 Dbh 0.76 0.58 43 Table 6. Table of variables and the number of regressions in which they were included in the final set of x variables. 44 Variable density percent mid Regressions densi high ty percent Crown area Crown volume Crown diameter Total softwood basal area Percent softwood in total basal area Total hardwood basal area Percent of hardwood in total basal area Crown closure Crown length Height Dbh Age Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal Basal area area area area area area area area area area area area area area area area area area area area area area area area area area area area area area of of of of of of of of of of of of of of of of of of of of of of of of of of of of of of overstory overstory overstory overstory overstory overstory overstory overstory overstory overstory overstory overstory overstory overstory understory understory understory understory understory understory understory understory understory understory understory understory understory understory understory understory balsam fir hemlock red spruce white pine white spruce big tooth aspen american beech amercian elm grey birch hop hornbeam paper birch red maple sugar maple yellow birch balsam fir cedar hemlock red spruce white spruce white ash N X X X big tooth aspen american beech american elm grey birch hop hornbeam paper birch red maple sugar maple yellow birch white pine X ACKNOWLEDGEMENTS We would like to thank the following people for field and laboratory assistance: M. Ambrose, J. Brushwein, J. Knight, J. Krall, T. Morrison, 8. Oliveri, D. Rounds, J. Rounds and H. Tedford. We would like, also, to thank Dr. Fred Stehr, Department of Entomology, Michigan State University, and Dr. John Witter, School of Natural Resources, University of Michigan for critical manuscript review. We would like to thank Dr. Robert Jensen of the College of Business Administration, University of Maine for his assistance with the statistical analyses. Funding of this project was made possible by a grant from the Maine Department of Conservation, Appropriation Number 1505.4011. 45 REFERENCES CITED Brown, W. L., T. Eisner, and R. H. Whittaker. 1970. Allomones and kairomones: Trans-specific chemical messengers. BioScience 20:21-2. Dixon, W. J., and F. J. Massey. 1969. Introduction pg Statistical Analysis. McGraw Hill, Inc. 638 pp. Dodge, H. R. 1961. Parasitism of spruce budworm by Glypta and Apanteles at different crown heights in Montana. Can. Entomol. 93:222-8. Doutt, R. L. 1959. Biology of parasitic Hymenoptera. Ann. Rev. Entomol. 4:161-82. Doutt, R. L. 1964. Biological characteristics of entomophagus adults, p. 145-67. £2.P° DeBach, ed., Biological Control p£_Insect Pests EE§.EEES§: Reinhold Pub. Co., New York. 844 pp. Dowden, P. B., and V. M. Carolin. 1950. Natural control factors affecting the spruce budworm in the Adirondacks during 1946-1948. J. Econ. Entomol. 43:774-83. Dowden, P. B., W. D. Buchanan, and V. M. Carolin. 1948. Natural control factors affecting the spruce budworm. J. Econ. Entomol. 41:457-64. Eyre, F. H. 1954. Forest cover types of the eastern United States. Report of the Committee on Forest Types, Society of American Foresters, washington. 67 pp. 46 47 Force, D. C. 1972. r- and k-strategists in endemic host parasitoid communities. Bull. Entomol. Soc. Am. 18:135—7. Hendry, L. B., P. D. Greany, and R. J. Gill. 1973. Kairomone mediated host-finding behavior in the parasitic wasp Orgilus lepidus. Entomol. Exp. Appl. 16:471-7. Jaynes, H. A., and A. Drooz. 1952. The importance of parasites in the spruce budworm infestation in New York and Maine. J. Econ. Entomol. 45:1057-61. Jones, R. L., W. J. Lewis, M. Beroza, B. A. Bierl, and A. N. Sparks. 1973. Host-seeking stimulants (kairomones) for the egg parasite, Trichogramma evanescens. Environ. Entomol. 2:593-596. Laing, J. 1937. Host finding by insect parasites. 1. Observations of the finding of hosts by Alysia manductor, Mormoniella vitripennis and Trichogramma evanescens. J. Anim. Ecol. 6:298-317. Laing, J. 1938. Host finding by insect parasites. J. Exp. Biol. 15:281-302. Leonard, D. E., and G. A. Simmons. 1974. The effects of Zectran on the parasitoids of the spruce budworm Choristoneura fumiferana (Lepidoptera: Tortricidae). Can. Entomol. 106:545-54. Leonard, D. E., G. A. Simmons, and G. K. VanDerwerker. 1973. Spruce budworm: Techniques to improve counting of eggs. J. Econ. Entomol. 66:992. Leopold, A. 1933. Game Management. Charles Schribner's Sons, New York, N.Y. 481 pp. 48 Lewis, W. J., R. L. Jones, and A. N. Sparks. 1972. A host-seeking stimulant for the egg parasite Trichogramma evanescens: its source and a demonstration of its laboratory and field activity. Ann. Entomol. Soc. Am. 65:1087-9. Lewis, W. J., R. L. Jones, D. A. Nordlund, and A. N. Sparks. 1975a. Kairomones and their use for management of entomophagous insects. I. Utilization for increasing rate of parasitism by Trichogramma spp. J. Chem. Ecol. 1:343-7. Lewis, W. J., R. L. Jones, D. A. Nordlund, and H. R. Gross, Jr. 1975b. Kairomones and their use for management of entomophagous insects. II. Principles causing increases in parasitization rates by Trichogramma spp. J. Chem. Ecol. 1:349-60. McGugan, B. M., and J. R. Blais. 1959. Spruce budworm parasite studies in northwestern Ontario. Can. Entomol. 91:758-93. Miller, C. A. 1953. Parasitism of spruce budworm eggs by Trichogramma minutum. Can. Dept. For. Entomol. and Pathol. Br., Bi-monthly Prog. Rep. 9:1. Morris, R. F. 1955. The development of sampling techniques for forest insect defoliators, with particular reference to spruce budworm. Can. J. 2001. 33:225-94. Morris, R. F., ed. 1963. The Dynamics g£_Epidemic Spruce Budworm Populations. Entomol. Soc. Can. Mem. 31:1-332. Mott, D. G. 1963. The forest and the spruce budworm, p. 189-202. £2_ R. F. Morris, ed., The Dynamics gf_Epidemic Spruce Budworm Populations. Entomol. Soc. Can. Mem. 31:1-332. 49 Muesebeck, C. F. W., H. K. Krombein, and H. K. Townes. 1951. Hymenoptera g£_America North of Mexico. U.S. Govt. Printing Off., Washington, D.C. 1040 pp. Nie, N. B., C. H. Hull, J. G. Jenkins, K. Steinbrenner, and D. H. Bent. 1970. Statistical Package for the Social Sciences. McGraw-Hill Co. 675 pp. Nordlund, D. A., W. J. Lewis, R. L. Jones, and H. R. Gross, Jr. 1976. Kairomones and their use for the management of entomophagus insects. IV. Effect of kairomones on productivity and longevity of Trichogramma pretiosum Riley (Hymenoptera: Trichogrammatidae). J. Chem. Ecol. 2:67-72. Price, P. W. 1973a. Parasitoid strategies and community organization. Environ. Entomol. 2:623-6. Price, P. W. 1973b. The development of parasitoid communities. Proc. Northeastern Forest Insect Work Conf. 5:29-41. Shahjahan, M., and F. A. Streams. 1973. Plant effects on host finding by Leiphron pseudopallipes (Hymenoptera: Braconidae), a parasitoid of the tarnished plant bug. Environ. Entomol. 2:921-5. Simmons, G. A. 1973. Conversion table for spruce budworm sampling units. Misc. Rep. Me. Life Sci. Agr. Exp. Stn., No. 151. 4 pp. Simmons, G. A., D. E. Leonard, and C. W. Chen.- 1975. Influences of tree species density and composition on parasitism of spruce budworm. Environ. Entomol. 4:832—6. Spurr, S. H. and B. V. Barnes. 1973. Forest Ecology. Ronald Press Co., New York, N.Y. 2nd Edition. 571 pp. 50 Sweetman, H. L. 1958. The Principles of Biological Control: Inter- relations g§_Hosts and Pests and Utilization in_Regulation of Animal and Plant Populations. Wm. C. Brown Co., DuBuque, Iowa. 560 pp. Thomas, H. A. 1966. Parasitism by Trichogramma minutum (HymenOptera: Trichogrammatidae) in the spruce budworm outbreak in Maine. Ann. Entomol. Soc. Am. 59:723-5. Wellington, W. G. 1958. The light reactions of the spruce budworm, Choristoneura fumiferana Clemens (Lepidoptera: Tortricidae). Can. Entomol. 80:56—82. Whittaker, R. H., and P. P. Feeny. 1971. Allelochemics: Chemical interactions between species. Science 171:757-70. Wilson, L. F. 1963. Host preference for oviposition by the spruce budworm in the Lake States. J. Econ. Entomol. 56:285-8. Wilson, L. F. 1964a. Oviposition site of the spruce budworm, Choristoneura fumiferana, modified by light. Ann. Entomol. Soc. Am. 57:643-5. Wilson, L. F. 1964b. Observation on the geo-orientation of spruce budworm, Choristoneura fumiferana, adults. Ann. Entomol. Soc. Am. 57:645-8. Wilson, L. F., and J. L. Bean. 1963. Site of spruce budworm masses on their preferred hosts in the Lake States. J. Econ. Entomol. 56:574-8. CHAPTER 3 THE INFLUENCE OF STAND FACTORS ON POPULATION LEVELS AND DISPERSAL LOSSES OF INSTAR-I AND -II SPRUCE BUDWORM ABSTRACT Data included intensive information on stand density, compo- sition, age, crown closure, crown surface area, basal area, height and associated vegetation across a 60 plot continuum in Maine. T-test results showed significantly higher numbers of budworm in the upper crowns when compared to the mid crowns in samples taken in the fall, and spring (PE0.0S). Survival rates of spring dispersing budworm were significantly lower in the high crown areas (pr.05). Regression analyses indicated that crown closure and non-host species in the understory and overstory had significant impact on survival rates of spring dispersing budworm. 51 INTRODUCTION The character of the forest has long been recognized as a fun- damental component in the development of spruce budworm (Choris- toneura fumiferana Clem.) outbreaks (Bakuzis and Hansen 1965, Basker- ville l975a, Batzer 1969, Blais 1968, 1973, Brown et a1. 1976, Craig- head 1923, Graham 1956, Graham and Knight 1965, Morris 1958, 1963, Morris and Bishop 1951, Mott 1963, Swaine et a1. 1924, Tothill 1923, VanRaalte 1972). Studies of spruce budworm population dynamics have indicated that stand density and composition influence larval survival (Batzer 1976, Mott 1963). Mott (1963) pointed out two general relation- ships apparent in forest/budworm interactions: (1) the effect of the forest on the budworm and (2) the effect of the budworm on the forest. This reports the influence of the forest on the spruce budworm. Previous studies indicated that high stand densities were con- ducive to survival of dispersing first- and second-instar budworms, while Open stands promoted large larval survival due to increased crown exposure (Batzer 1976). Mott (1963) postulated that maximum larval survival occurs where the product of large and small larval survival is maximum brought about by mature fir stands having well developed crowns without a significant component of non-host trees :or unstocked Openings. Studies leading to these postulates, however, «did not include actual measurements of dispersal losses immediately 52 53 following hatch or spring emergence from overwintering hibernaculae. Our work stressed the effects of various stand densities and com- positions on dispersing first- and second-instar spruce budworm. Measurements included actual survival rates of first- and second- instars in various stands representing a broad range of stages in spruce-fir forest succession. METHODS This study was conducted in central Aroostook and eastern Penob- scot counties of Maine (Leonard and Simmons 1974). Balsam fir (Abies balsamea (L.) Miller) predominated and other softwood species associ- ated were white spruce (Picea glauca (Muenchhausen Voss)), red spruce (P, rubens Sargent), hemlock (Tsuga canadensis (L.) Carr), white pine (Pinus strobus L.) and northern white cedar (Thuja occidentalis L.). Hardwoods consisted of yellow birch (Betula alleghaniensis Britton), white birch (B. papyrifera Marshall), grey birch (B, populifolia Marshall), sugar maple (Acer saccharum Marshall), red maple (A, rubrum L.), american elm (Ulmus americana L.), white ash, (Fraxinus americana L.), big tooth aspen (Populus grandidentate Michaux), american beech (Fagus grandifolia Ehrhart) and hop hornbeam (Ostrya virginiana (Miller) Koch). Sixty sites were chosen for date collection. An attempt was made to sample a continuum of forest types from nearly pure hardwood to pure softwood. Since balsam fir is the primary host for budworm, all sampled trees were balsam fir. Plots were 2 chains in diameter (40.2 m), with centers assigned to a single, mature balsam fir. To sample a Continuum, plots included 1-4 additional balsam fir trees. These additional (secondary) trees were on bearings north, south, east, and west of the center tree. The continuum sampled ranged from 1 primary 54 METHODS This study was conducted in central Aroostook and eastern Penob- scot counties of Maine (Leonard and Simmons 1974). Balsam fir (Abies balsamea (L.) Miller) predominated and other softwood species associ- ated were white spruce (Picea glauca (Muenchhausen Voss)), red spruce (P, rubens Sargent), hemlock (Tsuga canadensis (L.) Carr), white pine (Pinus strobus L.) and northern white cedar (Thuja occidentalis L.). Hardwoods consisted of yellow birch (Betula alleghaniensis Britton), white birch (B, papyrifera Marshall), grey birch (B, populifolia Marshall), sugar maple (Acer saccharum Marshall), red maple (A, rubrum L.), american elm (Ulmus americana L.), white ash, (Fraxinus americana L.), big tooth aspen (Populus grandidentate Michaux), american beech (Fagus grandifolia Ehrhart) and hop hornbeam (Ostrya virginiana (Miller) Koch). Sixty sites were chosen for date collection. An attempt was made to sample a continuum of forest types from nearly pure hardwood to pure softwood. Since balsam fir is the primary host for budworm, all sampled trees were balsam fir. Plots were 2 chains in diameter (40.2 m), with centers assigned to a single, mature balsam fir. To sample a continuum, plots included 1-4 additional balsam fir trees. These additional (secondary) trees were on bearings north, south, east, and west of the center tree. The continuum sampled ranged from 1 primary 54 55 balsam fir surrounded by non-host species to 1 primary balsam fir with 4 secondary fir trees at 4 points of the compass. These plots corresponded to predominantly hardwood and predominantly softwood areas respectively. Such plot selection functioned to provide forest suc- cessional stages for study. Earlier, more diverse ecological conditions were represented by plots containing 1 central balsam fir tree sur- rounded by non—host hardwoods. Later, less diverse successional stages were represented by plots containing almost entirely balsam fir, the sub-climax or climax species in the northern forest type of Maine (Eyre 1954). Spruce budworm population levels were measured three times during the study. In the late summer (August) of 1976, egg masses were collected in all plots to estimate initial population densities. The overwintering budworm population was measured by sampling after the first-instars had dispersed and spun hibernaculae (October-November, 1977). Spring budworm population levels were estimated by sampling after second-instar spring dispersal. In June, 1978, samples taken at the peak of the third-instar enabled estimates of established feeding population densities. Sampling in this manner enabled us to examine budworm population levels after each dispersal phase (fall and spring). As a result, survival rates (S) and mortality rates (1-S) were estimated for each dispersal period and compared to the vege- tational continuum studied. In the fall population sample, full branches were removed at two levels, high crown and mid crown, from both central and secondary sample trees. Five sample branches were taken from each level, 56 measured for length and width and transported to the lab for study. In the lab, branches were cut up and examined twice for egg masses. Cross-tabulation of egg mass length with the numbers of egg rows was used to derive the number of eggs per egg mass (Leonard et a1 1973). The number of viable eggs from the fall sample served to estimate the initial budworm population. Data collected from the sample branches consisted of egg masses, viable eggs, non-viable eggs, totals for each branch, and the total for the five branches at each sampling level. The nearly 4,000 egg masses from 1,890 branches contained 57,568 eggs. Populations during the overwintering stage were sampled in the same way. In the lab, branches were washed and larvae flushed and counted using a technique outlined by Miller et a1 (1971). Approximately 1,890 branches containing 69,445 second-instars were examined. Instead of full branch samples, 18 inch tip samples (56 cm) were collected and examined in the spring samples. Since budworms exhibit a positive phototaxis in the third instar, they tended to concentrate at the tips of balsam fir branches (Morris 1963). To compensate for error induced by sampling only at the tips, population values at the mid crown level were considered to be from the mean branch area of the mid crown overwintering samples. The same was done for the high crown samples. Approximately 1,890 branches were sampled containing 9,340 third-instars. There was a great deal of variability in branch size. Since the crowns of balsam fir tend to be conical, high crown branches tended to be smaller than mid crown branches. Therefore, all observations were 57 standardized to 9.3 m2 of balsam fir foliage (Simmons 1973). Stand data were collected in all plots. Basal areas of all tree species (understory and overstory) were measured in each prism plot (Avery 1975), using the primary balsam fir tree as a plot center. Data from primary and secondary balsam fir trees also consisted of height, dbh, crown length, crown diameter, crown closure, age, crown volume, and crown area. To determine if our proposed plot selection criteria had provided an ecological continuum of increasing balsam fir, a hierarchical clustering algorithm was used to analyze the data (Johnson 1967, Rohlf 1970). A cluster analysis was conducted on the basal area of balsam fir (overstory). Clustering in this way indicated the succes- sional continuum of a high basal area of balsam fir in some plots to a low basal area of balsam fir in other locations. T-test analyses and stepwise multiple linear regressions were used with both the upper and mid crown data to identify possible con- tributing factors to differences in survival rates. This was done for both fall and spring dispersal periods. One regression was run at the upper and mid crown level for all primary trees (60) using the respective survival rate for that crown height as the dependent variable. Independent variables included data from each sample tree such as height, dbh, crown diameter, crown closure, age, crown volume, and crown area. Other independent variables used were stand related factors such as basal areas of all tree species found in the plot. These basal areas were separated into 2 groups, overstory (dominant 58 or co—dominant) and understory (sub-dominant or repressed) trees. For example, 1 variate was overstory sugar maple and another variate was understory sugar maple. The least squares method of regression analysis was extended to a stepwise, forward inclusion, multiple regression model as described by Nie et a1 (1970) and Dixon and Massey (1969). Regressions were cal- culated using the Statistical Package for the Social Sciences with de- fault values of (n, f, t) converging on models of about 30 (x) variables (Nie et a1 1970). Therefore, the inclusion value was set at a maximum of 20 (x) variables. The P value and tolerance limit were commanded to be 1.00 and 0.01 respectively. RESULTS AND DISCUSSION Comparison of High and Mid Crown Densities In each sample, fall, winter, and spring, the mean number of insects was significantly greater in the high crown areas when compared with mid crown areas (P<0.05) (Table l). The significantly greater budworm levels in the high crowns of balsam fir in the fall sample indicated an oviposition preference exhibited by female budworm moths. This is consistant with several other authors who agree that female budworm moths tend to deposit eggs on foliage of taller spruce and fir trees due to an apparent positive phototaxis (Mott 1963, Wellington 1948, Wilson 1963, Wilson 1964a, Wilson 1964b, Wilson and Bean 1963). The higher budworm levels in the high crown areas during winter and spring suggest several relationships. The high crown may provide more suitable overwintering sites for budworm. Previous studies have shown that crown exposure influences radiation intensity and thus favorably affects body temperature of budworm at both dispersal times (Shepherd 1958, 1959). Spurr and Barnes (1973) implied that high crown areas receive the greatest amount of sunlight with mid and low crowns receiving much less light. Therefore, the higher numbers of budworm in high crown areas indicate a positive phototaxis on the part of the instars-I and-J]; This is adaptively significant. As a balsam fir tree ages, the micro-habitat associated with the tree changes. Young, 59 6O smooth bark becomes fissured and flaky with maturity (Westveld 1954). Physiological maturity also results in the production of staminate flowers (Blais 1959, Greenbank 1963a) which leave tiny cup-like bracts upon disintegration. Our observations also indicate that extensive lichen mats develop on the trunk and needle-free portions of the branches as balsam fir matures. Previous studies have shown that bark fissures, staminate flower bracts and lichen mats are commonly used by budworm larvae to spin overwintering hibernaculae (Miller 1958). Further, buds which have been excavated by previous generations of budworm are also used as overwintering sites (Batzer 1969). Other possible reasons for high budworm numbers in upper crowns include the effects of wind. Wind turbulence at the high crown level may keep the eclosing budworm in the high crown areas, thus minimizing percloation to lower portions. In mid crown areas, wind turbulence is greatly reduced from crown closure and proximity to the ground. There- fore, movement due to wind dispersal is less in mid crowns. Analyses showed that mean survival rates during the fall were greater and 1.0 and were not significantly different in the high and mid crowns (P>0.05) (Table 2). These greater than unit survival rates indicated an immigration to balsam fir. Jaynes and Speers (1949) showed that, in laboratory cages, white spruce was highly preferred for oviposition by budworm moths over balsam fir, red spruce, and black spruce. Wilson (1963) also found that in laboratory and field tests white spruce was preferred for oviposition over balsam fir by a ratio of over 2:1. Therefore, the immigration to balsam fir, observed 61 in this study may have been due to dispersal from other hosts such as white spruce. Staminate flower bracts on balsam fir are utilized by overwintering budworm larvae (Jaynes and Speers 1949, Miller 1958). The dispersal to balsam fir from other hosts, though not directed, is important in that preferred overwintering sites are available and over- wintering success is increased. The results, however, conflict with previous studies indicating significant fall dispersal mortality due to instar-I dispersal to non- host material (Miller 1958, 1975). These differences were probably due to sampling techniques of the two studies. Miller (1958, 1975) used emergence techniques to sample instar-II budworm. This study used an accurate extraction method of washing the balsam fir branches and collecting larvae (Miller 1971). Further, observations made by McKnight (1968) demonstrated that budworm were able to overwinter on non-host material. It follows that these larvae thought to be missing could disperse from the non-host material and thus contribute to the budworm population in the spring. The fall survival rates do not reflect naturally occurring mortality, but they indicate a net move- ment towards balsam fir from other hosts. These survival rates suggest that mortality during the fall dispersal may not be as significant as previously reported (Miller 1958, 1975). This is plausible since the larvae do not feed during the fall (Mott 1963). Survival rates of instar-II budworm dispersing in the spring, however, were much different. These values indicated the success of larvae in each crown level from overwintering instar-II budworm to establish feeding instar-III. Data showed survival rates significantly 62 greater in the mid crown areas when compared with high crown areas (P<0.05) (Table 2). The mean survival rates of mid and high crown areas indicated significant mortality during spring dispersal. These values were consistent with the percentage mortality reported by Miller (1958, 1975). There are a number of explanations for these differing rates of mortality from the fall to spring dispersal periods. Probably the most important factor is the need for the budworm to feed in the spring. Second—instars emerging from hibernaculae have a finite (internal) energy source available to them. The budworm, therefore, must find a food source shortly after emergence or they will perish. Recent studies have indicated that most emerging instar-II budworm die within 2 days at 26°C if they do not feed (Leonard, unpublished). It follows that if larvae overwinter on or disperse to non-host trees, they have a very short time to contact a host in order to survive. The differing survival rates in the high and mid crown regions are also worthy of consideration. Significantly greater mortality in the high crown areas of balsam fir may be due in part to the greater fluctuations in micro-climate associated with high crown areas (Spurr and Barnes 1973). Temperature fluctuations together with increased exposure to winds could serve to increase mortality of instar-II budworm. Greater wind activity in high crown areas increases the probability of high crown larvae dispersing to and dying on unsuitable vegetation. Though tree height is important, stand stocking densities should also be considered. Evidence indicated that by decreasing crown closure or interspersing stands with non-host material, those high 63 crown budworm more prone to dispersal have a greater chance of reaching unsuitable feeding sites (Mott 1963). Influences of Stand Factors As indicated earlier, fall dispersal survivals were generally greater than unit, indicating an immigration to balsam fir. To use regression analysis to attempt to infer forest component effects on survival of dispersing fall budworms, therefore, would be of no value. More information is needed in order to predict the actual survival rates in the fall (Tables 4, 5). Therefore, regression analyses were only used on the spring high crown and mid crown data to examine the effects of stand factors on spring budworm survival (Tables 3, 4). The period in the life cycle from spring emergence to establishment of feeding sites is very critical to the budworm. During this time, the individual budworm may undergo several dispersals. Each time the budworms disperse, there is a chance that they will fall through the canopy and be unable to return or land on an unsuitable non-host plant. Also during this time, the budworms are restricted in the energy they may expend. Since the budworms must establish feeding sites on a suitable host within a finite number of degree days, they are limited in the number of dispersals they can undergo. Those dispersing budworms that do not find their way to suitable feeding sites (current year shoots) soon perish. By increasing available areas unsuitable for budworms, such as non-host or open areas, it follows that the numbers of budworms lost to these areas would increase. 64 This would result from increased probability that dispersing larvae would find unsuitable feeding sites. Crown closure was positively correlated with spring budworm survival in both mid and high crowns (Tables 3, 4), indicating that as the canOpy closes, budworm survival increases. This is important in regulating wastage of larvae to the ground during spring dispersal. It follows that by decreasing crown closure, budworm survival would decrease due to greater losses of larvae to the forest floor. Decreased crown closure also encourages small openings on the forest floor which are known to support a wide variety of vegetation and a great diversity of habitat (Leopold 1933). This would positively affect the budworm predators and parasites by increasing alternate host habitats (Kemp and Simmons, in press; Simmons et a1 1975). Therefore, regulation of crown closure provides a practical method of managing the forest to induce spruce budworm mortality. Tables 3 and 4 indicate that several variables were negatively correlated with spring budworm survival. The spring mid crown re- gression analyses showed that understory red maple, understory red spruce, overstory beech, and overstory red maple were negatively correlated with spring budworm survival (Table 3). This suggests that by interspersing stands with these unsuitable budworm hosts, budworm survival would be lowered. Overstory non-hosts do not provide the necessary feeding sites for dispersing budworm. As a result, budworms must either re-disperse or perish. Further, since these hardwoods have not flushed during the budworm dispersal times, they provide areas 65 where larvae readily fall to the forest floor. Here, the probability that the budworms would be able to return to the forest canopy, before starvation, is quite small. Understory non-hosts would affect dis- persing spring budworms similarly. Larvae that fall through the canopy are unable to re-disperse due to the lack of atmospheric turbulence at the understory levels (Spurr and Barnes 1973). Larvae then starve attempting to establish feeding sites on non-hosts. Intuitively, these points seem important with regards to management of the forest against the budworm. However, these hon-host variables are much more sensitive to site conditions and may not be manipulated as easily as crown closure. Dbh was positively correlated with budworm survival in the Spring mid crown regression (Table 3). Since Dbh is generally indicative of tree age, this variable suggests that budworm survival increases as the stand matures. Indeed this has been found to be true in the past. Other variables positively correlated with spring budworm survival in the mid crown regression were overstory elm and overstory aspen. These positive correlations were most likely due to several reasons. Non-host materials in the overstory may, in fact, be suitable over- wintering sites for budworms (McKnight 1968). Therefore, in the spring, dispersing budworms from these overstory non-hosts may locate in lower regions of balsam fir or other adjacent host species. Here, non-host species in some cases can actually aid survival by supplying over- wintering sites from.which budworms can re-disperse to suitable hosts in the spring. This contrasts with previous studies that reported 66 non-hosts serve only as a source of mortality for dispersing budworms both in the fall and spring (Batzer 1969, Miller 1958, 1975, Mott 1963). Another reason for positive correlations between non-hosts and budworm survival is that these two species, overstory aspen and overstory elm, though associated with the forest type, are infrequent. Therefore, these species may not directly affect budworm survival, but do indicate the types of stands that are conducive to budworm survival. Analysis of the spring high crown regression revealed similar results. Negative correlations between high crown spring budworm survival and understory yellow birch and understory red maple, the two strongest variables in the regression, indicated that dispersal to understory areas offers little chance of survival of the budworms landing on these two species (Table 4). This suggests the budworms dispersing to the lower regions of the forest or forest floor likely will not survive, especially if non-hosts predominate the understory. A negative correlation between overstory aspen and budworm survival in the spring indicated that as overstory aspen increases, budworm survival decreases (Table 4). This results because aspen is unsuitable for food and bud- worms are lost to the ground during dispersal. Since aspen have not yet flushed when the budworms are dispersing, it follows that lack of interception allows more budworms to fall to the forest floor where chances of re-dispersing are small. The negative correlation between spring budworm survival and tree height indicated that conditions in the high crown were more environ- mentally severe than in the lower regions of the tree (Kemp and Simmons, in press) (Table 4). Therefore, higher crown areas should experience 67 the most drastic reductions in survival. Table 4 also shows that crown diameter of balsam fir was positively correlated with high crown spring budworm survival. As crown closure increases, crown diameter increases. Essentially, crown diameter acts similarly to crown closure in that with larger balsam fir crown diameters, more dispersing larvae are intercepted before falling to non-hosts or to the forest floor. It follows, therefore, that budworm survival increases with larger balsam fir crown diameters. Table 4 suggests that there are positive correlations between non-hosts and budworm survival. Overstory cedar, overstory ash, and overstory elm were all positively correlated with spring high crown budworm survival. These correlations indicated that, again, non-hosts were providing suitable overwintering sites from which budworm could disperse to suitable hosts in the spring. Though these were weak variables in the regression, they indeed aided budworm survival. Overstory hardwoods, on the other hand, can both reduce or aid survival. Overstory non-hosts can act to reduce budworm survival by increasing unsuitable food sources from which budworm must re-disperse in order to survive. Further, since overstory non-host material lacks foliage during budworm dispersal, greater numbers of budworms can be lost to the forest floor due to the lack of interception. This is much different than interception rates of conifers which hold their needles. Overstory non-hosts, however, can also aid spring budworm dispersal survival in that they provide added suitable overwintering sites from which budworms could re-disperse in the spring. 68 Understory non-host hardwoods decrease survival in similar ways. Larvae that fall through the canopy to a non-host understory, find non- host material unsuitable to feed on. However, in contrast to those budworms on overstory non-hosts, budworms in the understory cannot re-disperse due to the severe reduction in atmospheric turbulence in the understory (Spurr and Barnes 1973). Since understory non-hosts have not flushed during budworm dispersal, larvae can continue to fall through to the ground and perish. Understory non-hosts, indeed can be sources of suitable overwintering sites for budworms. However, the chances of survival of budworms overwintering on understory non-host material is very low. As indicated earlier, severe reductions in atmospheric turbulence prevent budworms from re-dispersing from these understory non-hosts and result in starvation. Clearly, non-host material is important in reducing budworm survival in the spring. By interspersing non-hosts in the understory and overstory, again, the forest is moving away from a subclimax forest of spruce and fir. Both mid crown and high crown regressions indicate several per- tinent factors relating to forest succession. More specifically, it seems that crown closure (open area) may be the most important variable that man can use to reduce budworm survival in the spring. Creating open areas represents a movement of the forest from a subclimax or climax stage to a somewhat earlier stage of succession where diversity is increased from open areas. Increased losses of budworms due to wastage to the forest floor or understory non-hosts, as well as increased losses of budworms due to greater numbers of predators and parasites, are all effects of open areas (crown closure). 69 Forest Management Implications Mott (1963) defined susceptibility of a forest to spruce budworm outbreaks as (l) the intrinsic factors that affect the survival rates of resident populations and (2) the extrinsic factors associated with location that affect dispersal of air-borne populations into and out of a stand. Since, for the most part, extrinsic factors are important during the moth stage of the budworm, this study concentrated on the intrinsic factors such as stand composition affecting the survival rates of resident populations of budworm larvae. High susceptibility is directly related to the accumulation of large, contiguous areas of balsam fir. Early cutting recommendations advocated removal of such aggregations and a reduction of susceptible forest stands (Westveld 1946, McLintlock 1947). Several other workers developed criterea for identifying highly susceptible forest stands (Balch 1946, Morris and Bishop 1951). Vulnerability, or the probability of damage to a stand once the budworm is present, has been the subject of much study. Several workers addressing this concept developed management plans and risk-rating systems for cutting practices (Swaine et a1 1924, Graham and Orr 1940, Westveld 1946, Balch 1946, McLintlock 1948, 1949, Morris and Bishop 1951, Turner 1952, Westveld 1954, Bean and Batzer 1956, Graham 1956, Morris 1958, Hatcher 1960, Blais 1964). Such silvicultural manipu- lation to minimize spruce budworm damage remains a forest management ideal. The damage that occurs in any forest following a budworm infes- tation depends on the severity of the attack (Mott 1963). This, in 70 turn, depends on the susceptibility of the forest. Mott (1963) suggests that the degree of damage from a constant level of infestation differs in different forest conditions. The ability of forests to withstand attack depends on many features, some of which affect sus- ceptibility as well as vulnerability. The abundance of non-host species and open areas are two factors that contribute to decreased vulner- ability of a stand. These two factors may also reduce susceptibility to a certain extent. While there is no question that silvicultural manipulation can decrease vulnerability, evidence indicates that it may also be important in reducing stand susceptibility. Several authors have outlined the merits and the problems of silvicultural control of the spruce budworm (Baskerville 1975a, 1975b, 1976, Batzer 1976). In light of their work, the interactions of the budworm and the spruce-fir forest are thought to bring long-run stability to the forest. Periodic outbreaks of the budworm insure regeneration of balsam fir forests. This, in turn, is important to the survival of the budworm. However, Baskerville (1976) points out that though the budworm and the forest are environmentally stable, from man's point of view, the forest is not economically stable. This is due to losses of large acreages in a short period of time during budworm infestations. As Baskerville (1976) suggests, in order to maintain the present level of industry based on the spruce-fir forest, the forest must both be protected and actively managed. Regardless of what management practices are adopted, they will not be without problems (Batzer 1976). 70 turn, depends on the susceptibility of the forest. Mott (1963) suggests that the degree of damage from a constant level of infestation differs in different forest conditions. The ability of forests to withstand attack depends on many features, some of which affect sus- ceptibility as well as vulnerability. The abundance of non-host species and open areas are two factors that contribute to decreased vulner- ability of a stand. These two factors may also reduce susceptibility to a certain extent. While there is no question that silvicultural manipulation can decrease vulnerability, evidence indicates that it may also be important in reducing stand susceptibility. Several authors have outlined the merits and the problems of silvicultural control of the spruce budworm (Baskerville 1975a, 1975b, 1976, Batzer 1976). In light of their work, the interactions of the budworm and the spruce-fir forest are thought to bring long-run stability to the forest. Periodic outbreaks of the budworm insure regeneration of balsam fir forests. This, in turn, is important to the survival of the budworm. However, Baskerville (1976) points out that though the budworm and the forest are environmentally stable, from man's point of view, the forest is not economically stable. This is due to losses of largeacreagesin a short period of time during budworm infestations. As Baskerville (1976) suggests, in order to maintain the present level of industry based on the spruce-fir forest, the forest must both be protected and actively managed. Regardless of what management practices are adopted, they will not be without problems (Batzer 1976). 71 In view of the results of this report, more research is needed in areas of actual silvicultural manipulation of forest stands. Various types of management practices should be implemented. Patch and block clearcuts as well as selection cutting should be implemented and com- pared to unmanaged forests. Further, these practices should be evaluated over long periods of time so that the effects of various silvicultural maniuplations on the susceptibility and especially vulnerability could be continually assessed. Table l. 72 T-values and probabilities of spruce budworm viable eggs (instar-I), overwintering larvae (instar-II), and estab- lished spring larvae (instar-III) in comparing high and mid crown balsam fir samples (9.3 m2 foliage). Island Falls, Vanceboro, Maine, 1976. mmMIm N®.N© Nm.m® mmH HHH l Hmumcfl Ho.ovm .mem vm.v czouo be: ommuo va.mm Hm.moa mma HHH I Hmumca czouo so“: mcfiumm 73 NHmImH mm.mnm m®.th hmH HH l HmumCfl Ho.ovm .mem oa.m szono be: mmmalmm mm.mmv mh.mmv bma HH I nuance csouo swam kucflz mmmauo cm.0mm ma.nm~ was mmmm wands> mo.ovm .mflm mm.v czouo be: mmmm manmfi> mnmmuo vm.moo vm.0mm ems :3oso roam Hamm .noum HHMEIH czam>u9 mmcmm m mmsouo Ix z 74 Table 2. T-values and probabilities of spruce budworm survival rates during the fall (instar-I) dispersal and the spring (instar-II) dispersal in comparing high and mid crown balsam fir samples. Island Falls, Vanceboro, Maine, 1976. 75 mo.ova .mam om.o Hm.a Hm.o hm.o mm.o 00 00 czono pee mums Hm>fi>H5m HH I umumcH cacao noes mums Hm>fl>hsm HH I umumCH ammuomwfio mcflumm mo.OAa .mam: mm.H Hm.H mH.N ov.H mm.H vm om csouo wee wumn Hm>flbhsm H I HmumcH c3ouo no“: oumn Hm>w>usm H I HmumcH ammuwmmflo Hamm .oflm Hameua oDHm>IB m wmsouu 76 Table 3. Variables included in the final mid crown regression equation using survival rates of spring dispersing instar-II budworm (y) and stand parameters (x). Island Falls, Vanceboro, Maine, 1976. 77 Variable B R2 Overstory aspen 0.019 0.10 Crown closure 0.008 0.16 Overstory elm 0.063 0.21 Understory red maple -0.593 0.24 Understory red spruce -0.l32 0.28 Overstory beech -0.026 0.32 Dbh 0.069 0.37 Overstory red maple -0.015 0.40 Total basal area -0.003 0.42 78 Table 4. Variables included in the final high crown regression equation using survival rates of spring dispersing instar-II budworm (y) and stand parameters (x). Island Falls, Vanceboro, Maine, 1976. 79 Variables B R2 Overstory cedar 0.004 0.12 Height -0.010 0.17 Crown closure 0.006 0.22 Crown diameter 0.022 0.26 Understory yellow birch -0.892 0.30 Overstory aspen -0.006 0.36 Overstory ash 0.028 0.41 Understory red maple -0.374 0.44 Overstory elm 0.022 0.48 80 Table 5. Variables included in the final mid crown regression equation using survival rates of fall dispersing instar-I budworm (y) and stand parameters (x). Island Falls, Vanceboro, Maine, 1976. 81 Variable B R2 Overstory sugar maple 0.121 0.11 Overstory ash -0.427 0.19 Dbh -0.581 0.24 Understory yellow birch 8.676 0.33 Height 0.026 0.40 Understory red maple 1.569 0.45 Understory balsam fir -0.252 0.48 Overstory balsam fir 0.039 0.52 82 Table 6. Variables included in the final high crown regression equation using survival rates of fall dispersing instar-I budworm (y) and stand parameters (x). Island Falls, Vanceboro, Maine, 1976. 83 Variable B R2 Overstory beech 0.504 0.13 Dbh -0.555 0.30 Understory yellow birch -15.247 0.40 Overstory paper birch -0.65 0.45 Understory cedar -0.275 0.48 Overstory red spruce 0.035 0.50 Height -0.049 0.52 Crown length 0.075 0.54 Overstory red maple 0.077 0.57 ACKNOWLEDGEMENTS We would like to thank Dr. Daniel T. Jennings, Principal Insect Ecologist, Northeast Forest Experiment Station, USDA, Forest Service for assistance and direction during the field study phase. We would also like to thank the following people for field and laboratory assistance: M. Ambrose, J. Brushwein, J. Knight, J. Krall, T. Morrison, S. Oliveri, D. Rounds, J. Rounds, H. Tedford, and J. Walker. Very special thanks to Cheryl Kemp. Funding of this project was made possible by a grant from the Maine Department of Conservation, ApprOpriation number 1505.4011. 84 REFERENCES CITED Avery, T. E. 1975. Natural resources measurements. McGraw-Hill Co. 339 pp. Bakuzis, E. V., and H. L. Hansen. 1965. Balsam Fir: Abies balsamea (L.) Mill. A Monographic Review. Univ. Minn. Press, Minneapolis. 445 pp. Balch, R. E. 1946. The spruce budworm and forest management in the Maritime Provinces. Can. Dep. Agr., Entomol. Div. Processed Pub. 60. 7 pp. Baskerville, G. L. 1975a. Spruce budworm: Super silviculturist. Forest Chron. 51:4-6. Baskerville, G. L. 1975b. Spruce budworm: The answer is forest management: or is it? Forest Chron. 51:23-26. Baskerville, G. (Task force leader) 1976. Report of the task force for evaluation of budworm control alternatives. Prep. For. Cab. Comm. Econ. Dev. Prov. New Brunswick. 205 pp. Batzer, H. O. 1969. The forest character and vulnerability of balsam fir to spruce budworm in Minnesota. Forest Sci. 15:17-25. Batzer, N. 0. 1976. Silvicultural control techniques for the spruce budworm. In_Chansler, J. F., and W. H. Klein, pgs. 110-116, Proceedings of a symposium on the spruce budworm, November 11-14, 1974, Alexandria, Virginia. USDA Misc. Pub. No. 1327. 188 pp. 85 86 Bean, J. L., and H. O. Batzer. 1956. A spruce budworm risk rating for the spruce fir types in the lake states. USDA For. Serv. Tech. Note 453, Lake States forest Exp. Sta. St. Paul, Minn. 2 pp. Blais, J. R. 1959. The vulnerability of balsam fir to spruce budworm attack in northwestern Ontario, with special reference to the physiological age of the tree. 34:405-422. Blais, J. R. 1964. Account of a recent spruce budworm outbreak in Laurentide park region of Quebec and measures for reducing damage in future outbreaks. Forest Chron. 40:313-323. Blais, J. R. 1968. Regional variation in susceptibility of eastern north American forests to budworm attack based on history of outbreaks. Forest Chron. 44:17-23. Brown, M. W., F. B. Knight and J. B. Dimond. 1976. Stand composition and susceptibility to spruce budworm epidemics. Univ. Me. Sch. Forest Res. Tech Note No. 61. 4 pp. Craighead, F. C. 1923. A brief summary of the budworm investigations in Canada. J. Forestry 21:135-138. Dixon, W. J., and F. J. Massey, 1969. Introduction to Statistical Analysis. McGraw-Hill, Inc. 638 pp. Eyre, F. H. 1954. Forest cover types of the eastern United States. Report of the committee on forest types, Society of American Foresters, Washington. 67 pp. Graham, S. A. 1956. Hazard rating of stands containing balsam fir according to expected injury by spruce budworm. Univ. Mich. Dep. Forest. Mich. Forest. 13. 2 pp. 87 Graham, S. A., and F. B. Knight. 1965. Principles of forest entomology. Ed. 4. McGraw-Hill, New York. 417 pp. Graham, S. A., and L. W. Orr. 1940. The spruce budworm in Minnesota. Univ. Minn. Agr. Exp. Sta., Tech. Bull. 142. 27 pp. Greenbank, D. O. 1963. Staminate flowers and the spruce budworm. In_ R. F. Morris (ed), Pgs. 208-218, The dynamics of epidemic spruce budworm populations. Entomol. Soc. Can. Mem. 31:1-32. Hatcher, R. J. 1960. Development of balsam fir following a clearcut in Quebec, Can. Dep. North. Affairs and Natu. Resourc., Forest Resour. Div., Tech. Note 87. 22 pp. Jaynes, H. A., and C. F. Speers. 1949. Biological and ecological studies of the spruce budworm. J. Econ. Entomol. 42:221-225. Johnson, S. C. 1967. Hierarchical clustering schemes. Psychometrica 36:241-254. Leonard, D. E., and G. A. Simmons. 1974. The effects of Zectran on the parasitoids of the spruce budworm Choristoneura fumiferana (Lepidoptera: Tortricidae). Can. Entomol. 106:545-554. Leonard, D. E., G. A. Simmons, and G. K. VanDerWerker. 1973. Spruce budworm: techniques to improve counting of eggs. J. Econ. Entomol. 66:992. McKnight, N. E. 1968. A literature review of the spruce, western, and 2-year-cycle budworms Choristoneura fumiferana, E, occidentalis, and g, biennis (Lepidoptera: Tortricidae). U.S.D.A. For. Ser. Res. Pap. RM-44 35 pp. McLintock, T. F. 1947. Silvicultural practices for control of the spruce budworm. J. Forestry 45:655-658. 88 McLintock, T. F. 1948. Evaluation of tree risk in the spruce-fir region of the northeast. Iowa State Coll. J. Sci. 22:415-419. McLintock, T. F. 1949. Mapping vulnerability of spruce fir stands in the northeast to spruce budworm attack. USDA Forest Serv. Sta. Pap. 21, northeast forest Exp. Sta., Upper Darby, Pa. 20 pp. Miller, C. A. 1975. Spruce budworm: how it lives and what it does. Forest Chron. 51:2-4. Miller, C. A., E. G. Kettela, and G. A. McDougall. 1971. A sampling technique for overwintering spruce budworm and its applicability to p0pulation surveys. Can. Forest. Rep. M-X-25. Fredericton, New Brunswick. 11 pp. Morris, R. F. 1958. A review of the important insects affecting the spruce-fir forest in the Maritime Provinces. Forest Chron. 34:159-189. Morris, R. F. 1963. Resume. In_R. F. Morris (ed.), pgs 311-320, The dynamics of epidemic spruce budworm populations. Entomol. Soc. Can. Mem. 31:1-332. Morris, R. F., and R. L. Bishop. 1951. A method of rapid forest survey for mapping vulnerability to spruce budworm damage. Forest Chron. 27:171-178. Mott, D. G. 1963. The forest and the spruce budworm. IE_R. F. Morris (ed.), pgs 189-202, The dynamics of epidemic spruce budworm populations. Entomol. Soc. Can. Mem. 31:1-332. Nie, N. H., C. H. Hull, J. G. Jenkins, K. Steinbrenner, and D. H. Bent. 1970. Statistical Package for the Social Sciences. McGraw-Hill Co. 675 pp. l“ 88 McLintock, T. F. 1948. Evaluation of tree risk in the spruce-fir region of the northeast. Iowa State Coll. J. Sci. 22:415-419. McLintock, T. F. 1949. Mapping vulnerability of spruce fir stands in the northeast to spruce budworm attack. USDA Forest Serv. Sta. Pap. 21, northeast forest Exp. Sta., Upper Darby, Pa. 20 pp. Miller, C. A. 1975. Spruce budworm: how it lives and what it does. Forest Chron. 51:2-4. Miller, C. A., E. G. Kettela, and G. A. McDougall. 1971. A sampling technique for overwintering spruce budworm and its applicability to population surveys. Can. Forest. Rep. M-X-25. Fredericton, New Brunswick. 11 pp. Morris, R. F. 1958. A review of the important insects affecting the spruce-fir forest in the Maritime Provinces. Forest Chron. 34:159-189. Morris, R. F. 1963. Resume. lE.R- F. Morris (ed.), pgs 311-320, The dynamics of epidemic spruce budworm populations. Entomol. Soc. Can. Mem. 31:1-332. Morris, R. F., and R. L. Bishop. 1951. A method of rapid forest survey for mapping vulnerability to spruce budworm damage. Forest Chron. 27:171-178. Mott, D. G. 1963. The forest and the spruce budworm. In_R. F. Morris (ed.), pgs 189-202, The dynamics of epidemic spruce budworm populations. Entomol. Soc. Can. Mem. 31:1-332. Nie, N. H., C. H. Hull, J. G. Jenkins, K. Steinbrenner, and D. H. Bent. 1970. Statistical Package for the Social Sciences. McGraw—Hill Co. 675 pp. 89 Rohlf, F. J. 1970. Adaptive hierarchical clustering schemes. Systematic Zoology. 19:58-82. Shepherd, R. F. 1958. Factors controlling the internal temperature of spruce budworm larvae, Choristoneura fumiferana (C1em.). Can. J. Zool. 36:779-786. Shepherd, R. F. 1959. Phytosociological and environmental characteristics of outbreak and non outbreak areas of 2-year cycle spruce budworm, Choristoneura fumiferana. Ecology 40:608-620. Simmons, G. A. 1973. Conversion table for spruce budworm sampling units Misc. Rep. Me. Life. Sci. Agr. Exp. Stn., No. 151. 4 pp. Spurr, S. H. and B. V. Barnes. 1973. Forest Ecology. Ronald Press Co. New York, N.Y. 2nd edition. 571 pp. Swaine, J. M., F. C. Craighead, and J. W. Bailey. 1924. Studies on the spruce budworm. Part II. General bionomics and possibilities of prevention and control. Can. Dep. Agr. Bull. 37. 91 pp. Tothill, J. D. 1923. Notes on the outbreaks of spruce budworm, forest tent caterpillar and larch sawfly in New Brunswick. Acadian Entomol. Soc. Proc. 8:172-182. Turner, K. B. 1952. The relation of mortality of balsam fir caused by spruce budworm to forest composition in the Algoma forest of Ontario Can. Dep. Agr. Pub. 187. 107 pp. VanRaalte, G. D. 1972. "Do I have a budworm-susceptible forest?" Forest Chron. 48:190-192. 90 Wellington, W. G. 1948. The light reactions of the spruce budworm, Choristoneura fumiterana Clemens (Lepidoptera: tortricidae) Can. Entomol. 80:56-82. Westveld, M. 1946. Forest management as a means of controlling the spruce budworm. J. Forest. 44:949-953. Westveld, M. 1954. A budworm vigor resistance classification for spruce and balsam fir. J. Forestry 52:11-24. Wilson, L. F. 1963. Host preference for oviposition by the spruce budworm in the lake states. J. Econ. Entomol. 56:285-288. Wilson, L. F. 1964a. Oviposition site of the spruce budworm, Choristoneura fumiferana, modified by light. Ann Entomol. Soc. Am. 57:643-645. Wilson, L. F. 1964b. Observation on the geo-orientation of spruce budworm Choristoneura fumiferana, adults. Ann. Entomol. Soc. Am. 57:645-648. Wilson, L. F., and J. L. Bean. 1963. Site of spruce budworm egg masses on their preferred hosts in the lake states. J. Econ. Entomol. 56:574-578. CHAPTER 4 A MODEL FOR EXAMINING THE EFFECTS OF STAND FACTORS ON SPRUCE BUDWORM LARVAL DISPERSAL ABSTRACT A stochastic model was developed to simulate the life history of in- star-I through instar-II spruce budworm (Choristoneura fumiferana (Clemens)). Particular emphasis was given to the dispersal phases during spring and fall and the factors affecting budworm survival. Analyses included the effects of aggregated stand types, temperatures, cloud cover, Open area and the species distribution of the host com- ponent on dispersal mortality. Of those stand factors examined, open area and non-host percentages were the most important in regulating spring dispersal survival. 91 INTRODUCT ION Few procedures have been developed to model larval dispersal in a heterogenous environment. Watt (1966) developed and Kitching (1971) used a system where inputs and outputs for each discrete habitat unit were re- corded and summed with iterations for each time interval in the dispersal period. For this case, a stochastic approach was desirable to model dis- persal within an aggregate essemblage of trees rather than as a process influenced by individual trees. With this approach, the effects of stand composition and spatial pattern on dispersal were investigated with great— er ease. Using the theory of the nearest neighbor measure of spatial re- lations, a model was developed, incorporating past and present information, to define effects of stand composition, spacing, and weather on the dis- persal losses of spruce budworm during the fall and spring. 92 METHODS Model Description - General Diagrams of hypothetical stands were constructed by assigning a nu- meric value to each block on a grid designating either a host or a non- host tree. For each tree in the grid pattern, the probability of encoun- tering a host or non-host as a nearest neighbor in a hexagonal arrangement was calculated (Clark and Evans 1954). Averaging these calculations for the stand gave four conditional probabilities: the probability of encoun- tering a host or non-host given that the starting point is a host and the same probabilities given that the starting point is a non-host. A generalized dispersal model was formulated by first assuming that the host species (black, white and red spruce, and balsam fir) were ran- domly dispersed throughout the areas designated as host. Open areas were distributed randomly throughout the stand. In addition, it was assumed that the probability of a larva hitting an open area was pr0portional to available open area. Finally, it was assumed that the probabilities of interception by a host or non-host, calculated by evaluating the nearest neighbors, were the same for all distances travelled by the dispersing larvae. Since most dispersal occurs from directionless updrafts, dispersal was simulated in random directions. A proportion of larvae were assigned to hosts, non-hosts and Open areas as a function of the probability of intercepting each of these after the larvae left a source. Mortality 93 94 occurred when larvae impacted Open areas. Twenty percent of the larvae intercepted by non-hosts were also assumed to eventually reach the forest floor and die. Prior to each dispersal, the larvae were lumped as a com- mon variable. A more specific description of fall and spring dispersal follows. Model Description - Specific In the fall and spring, the number of larvae dispersing is a propor- tion of the total available to disperse. A number of factors contribute to determining this proportion. Stand density and cloud cover influence both fall and spring dispersal. Larval dispersal occurs when the photo- positive larvae (Wellington 1948) crawl to the tips of the branches, spin silk threads, and are carried by the wind (Henson 1950). Two factors limiting dispersal are the number of larvae responding photopositively and turbulent wind conditions capable of launching larvae. First-instar budworm emerge from the egg mass over a two week period (Mott 1963). Cloud cover determines the actual number dispersing each day. In the model, the dispersal period was aggregated into one disper- sion. The effect of weather was simplified where the proportion of dis- persing larvae was dependent on the number of (clear) dispersal.days. This is a valid simplification of Shaw and Little (1973) who found vir- tually no dispersal occurring on overcast days, thus, weather effect was expressed as a percentage of favorable days during the dispersal period. Spring dispersal occurs repeatedly, depending on the length of time re- quired to accumulate a specified number of degree days. In this case, the proportional effect of weather was divided by the number of dispersals occurring. 95 The transfer of small larvae from tree to tree or from tree to open areas depends on air currents breaking silk threads spun by the larvae. A greater dispersing proportion would be expected in Open stands which permit free air movement facilitating updrafts. Morris and Mott (1963) found that approximately 50 percent of the larvae dispersed in dense stands and almost all dispersed in open stands. Since stand stocking in the model is held constant, stand density must be defined in terms of the non-host trees that are leafless and contribute to the "openness" of the stand. The multiplier of larvae dispersing as influenced by stand density is defined as: Y = .5 + (.6 x NHOST) + (1.2 x OPEN) (l)* NHOST and OPEN are the non-host and open components respectively. The proportion of larvae dispersing from the total available larvae, due to the effect of stand density and weather, is then: DISLAR = TOTLAR x Y x WEATH (2) DISLAR are the dispersing larvae, TOTLAR are the total larvae available, Y is the stand density factor with a range of 0.5 to 1.0 and WEATH is the proportion of cloud cover with a range of 0.0 to 1.0. During spring dis- persal WEATH is defined as: (WEATH) / (number of dispersals) Parasitism by Trichogramma minutum Riley, Apanteles fumiferanae Viereck and Glypta fumiferanae (Viereck) influence the number of disper- sing larvae. T, minutum is an egg parasite which reduces the total number of budworm available to disperse from the egg stage. E, fumiferanae and * Numbers correspond to indicated areas of the flow diagrams in Figures 1"30 96 A, fumiferanae, while not causing mortality in the first- or second-instars, alter the behavior of parasitized larvae and reduce the number dispersing. Parasitism by each species is dynamic and is influenced by stand com- position and density. Kemp and Simmons (in press) found that T, minutum parasitism rates rose with with increasing density of non-budworm host tree species. This was attributed to an increased availability of alter- native hosts for T, minutum. In the model, E, fumiferanae and A, fumiferanae also increase with increased levels of open areas and non- hosts. Simmons et a1 (1975) suggested that spruce budworm parasite den- sities are influenced by tree species density and composition. A more diverse forest with lower densities of spruce and fir provides a greater variety of alternate host insects. Openings in the canopy facilitate sup- port of herbaceous plants on the forest floor providing nectar, an impor- tant food of adult parasites (Syme 1966, 1975, Price 1976, Leuis 1967). In addition, these Openings retain the parasites within the stand, reduce energy expenditures needed for successful host exploitation and increase the chances of parasitism (Simmons et a1 1975). Using information present- ed by Kemp and Simmons (in press) T, minutum parasitism is given as: PAR = .01 + (NHOST x .58) (3) ‘g. fumiferanae and A, fumiferanae parasitism are given as: PAR = .1 + (.3 x NHOST) + (.55 x OPEN) ‘ (4) Ranges for this rate were obtained from Simmons et al (1977). g, fumiferanae and A, fumiferanae alter the behavior of parasitized larvae by changing the phototactic response from positive to negative (Wellington 1948). Lewis (1960) found that 25 percent of the parasitized larvae did not disperse. TOTLAR used in equation 2 was re-defined for spring dispersal as the unparasitized budworm plus 75 percent of the 97 parasitized budworm: TOTLAR = BUDW + (PARBUD x .75) ’ (5) where BUDW are the normal budworm and PARBUD are the parasitized budworm. PARBUD is determined: PARBUD = TOTBUD x PAR (6) and BUDW is: BUDW = TOTBUD - PARBUD (7) Where fall dispersal occurs only once in the model, spring dispersal takes place repeatedly. Two factors interact to determine the number of larvae dispersing after the affects of cloud cover, stand density, and parasitism have been evaluated--the preferences for each host exhibited by larvae and the number of dispersals. Shaw and Little (1973) reported that most budworm dispersal occurred during the warmest part of the day. Therefore, spring dispersal was modelled as taking place once each day over a length of time determined by temperature conditions. It was assumed dispersal would continue until larvae starved. Starvation results at 78 degree days from the start of dispersal (Leonard, unpublished). Thus, if the weather during the spring dispersal period was cool, more dispersals could occur than if it had been warm. Jaynes and Speers (1949) measured the percentage of larvae reaching expanding buds in the spring. Those larvae not reaching the bud sites were assumed to have dispersed. The average of two reported values each for balsam fir, red spruce and black spruce were 56, 27, and 11 percent respectively. The ordering of these percentages corresponds to the re- ported budworm preferences for these species (Blais 1957). These figures and an interpolated value for the white spruce of 30 percent were used as 97 parasitized budworm: TOTLAR = BUDW + (PARBUD x .75) ' (5) where BUDW are the normal budworm and PARBUD are the parasitized budworm. PARBUD is determined: PARBUD = TOTBUD x PAR (6) and BUDW is: BUDW = TOTBUD - PARBUD (7) Where fall dispersal occurs only once in the model, spring dispersal takes place repeatedly. Two factors interact to determine the number of larvae dispersing after the affects of cloud cover, stand density, and parasitism have been evaluated-the preferences for each host exhibited by larvae and the number of dispersals. Shaw and Little (1973) reported that most budworm dispersal occurred during the warmest part of the day. Therefore, spring dispersal was modelled as taking place once each day over a length of time determined by temperature conditions. It was assumed dispersal would continue until larvae starved. Starvation results at 78 degree days from the start of dispersal (Leonard, unpublished). Thus, if the weather during the spring dispersal period was cool, more dispersals could occur than if it had been warm. Jaynes and Speers (1949) measured the percentage of larvae reaching expanding buds in the spring. Those larvae not reaching the bud sites were assumed to have dispersed. The average of two reported values each for balsam fir, red spruce and black spruce were 56, 27, and 11 percent respectively. The ordering of these percentages corresponds to the re- ported budworm preferences for these species (Blais 1957). These figures and an interpolated value for the white spruce of 30 percent were used as 98 an indication of budworm preference. These preferences change with increasing total degree days--when the starvation point is reached, none of the larvae disperse. Third-instar budworm suffer greater mortality on black spruce due to late bud break (Blais 1957). In addition, larval development is slower (Blais 1957). Similar differences may occur on other host species and thus, the stimulus to redisperse may increase the chances of survival. At present, however, published data is not available so preference functions did not include this aspect. The preference function for all host species is given below: BS = .08 + (.00341 x TOTDD) RS = .27 + (.00270 x TOTDD) (8) WS = .30 + (.00259 x TOTDD) BF = .56 + (.00163 X TOTDD) The preference for the non-host trees is always 0. At the completion of dispersal, larvae on non-hosts and 50 percent of those on black spruce perish. Overwintering mortality as described by Miller (1958) was constant. While this is in no way indicative of a real world situation, no data is currently available with which to model a dynamic winter mortality. Finally, degree days were calculated each day using the method outlined by Baskerville and Emin (1969). Degree days were used to determine the length of spring dispersal. 99 Analysis This model was utilized to (1) bring together as much information about the spruce budworm as possible and (2) aid in the development of future forest management methods to control the spruce budworm. Since the spruce budworm is most vulnerable during fall and spring dispersals, it follows that information of this nature should be incorporated in future management techniques. In this model, there were three periods of mortality - fall, winter, and spring. Winter mortality was constant throughout the model. This is defined by Miller (1958) as such. However, mortality in the fall and spring were both dynamic and based on a number of variables such as stand composition, stocking, and weather. Analyses, therefore, emphasized the relationships involved in fall and spring dispersals. In order to effectively analyze the components of this model, three types of forests were established. These included non-host densities of approximately 9%, 41%, and 65%. Corresponding host densities were 81%, 49%, and 25%. Open area was considered to be randomly distributed in both the host and non-host components of the stand. By altering different components of the model, we were able to examine (1) the effects of clumped vs. random spacing of host and non-host species (2) the effects of tem- perature (3) the effects of cloud cover (4) the effects of open area and (5) the effects of host species composition on dispersal mortality. 99 Analysis This model was utilized to (1) bring together as much information about the spruce budworm as possible and (2) aid in the development of future forest management methods to control the spruce budworm. Since the spruce budworm is most vulnerable during fall and spring dispersals, it follows that information of this nature should be incorporated in future management techniques. A In this model, there were three periods of mortality - fall, winter, and spring. Winter mortality was constant throughout the model. This is defined by Miller (1958) as such. However, mortality in the fall and Spring were both dynamic and based on a number of variables such as stand composition, stocking, and weather. Analyses, therefore, emphasized the relationships involved in fall and spring dispersals. In order to effectively analyze the components of this model, three types of forests were established. These included non-host densities of approximately 9%, 41%, and 65%. Corresponding host densities were 81%, 49%, and 25%. Open area was considered to be randomly distributed in both the host and non-host components of the stand. By altering different components of the model, we were able to examine (l) the effects of clumped vs. random spacing of host and non-host species (2) the effects of tem- perature (3) the effects of cloud cover (4) the effects of open area and (5) the effects of host species composition on dispersal mortality. RESULTS AND DISCUSSION Figure 4 was the basis for all comparisons. Here spring dispersal mortality was greater than fall dispersal mortality. Further, spring dispersal losses showed different rates of mortality increase, where fall dispersal mortality showed a constant increase. These relationships remained true for all the variations tested. Figure 4 shows an increase in dispersal mortality in both the fall and the spring as non-hosts in- crease. Stand spacing was the first component analyzed. By aggregating hosts and non-hosts, we found that fall and spring budworm dispersal mortalities were slightly reduced (Figure 5). Figure 5 illustrates the increased success of finding suitable hosts when budworms originate from suitable hosts. Since temperature is considered an important factor in the spring dispersal process, the effects of a shOrter dispersal time due to warm weather were examined. In Figure 6 all variables were held constant except daily temperatures. In this case, maximum-minimum temperatures from a warmer spring were used. Fewer dispersals result from warmer springs. Therefore, there is a decrease in the losses of budworms from dispersal to unsuitable hosts in the spring. Figure 7 and 8 when compared with Figure 4, indicate the effects of cloudy weather on the dispersal losses. All components were held constant except that during fall and spring dispersals, 25% and 75% of the days 100 101 were cloudy. Figure 7 shows a slight reduction in dispersal mortality where Figure 8 shows a great reduction in spring dispersal mortality, when compared with Optimal cloud conditions (Figure 4). Figures 6-8 indicate environmental effects that cannot be controlled. The effects can, in some cases, increase budworm mortality in forest stands, however. The components that can be manipulated are very important in that they are the only means we have for managing spruce budworm with silvicultural techniques. As indicated earlier, Figure 4 shows that as non-hosts in a stand increase, the percent mortality during the fall and spring dispersals increase. Not only do the non-hosts directly affect mortality from the interspersion of unsuitable feeding sites, but they also, in conjunction with open areas, influence parasitism rates which are important in fall mortality and spring dispersal. Figure 4 shows the effect of 10% open area. By increasing the open areas to 20% (Figure 9) and 30% (Figure 10), we found that there were rather large increases in mortality. This was true especially in the spring and at lower densities of non-host species. Figure 11 shows that by reducing the open area in the stand to 1%, there was a marked reduction in fall and spring dispersal mortalities. These figures indicate that open area is one of the most important components that can be manipulated. Another forest component that can be manipulated is the species composition of the host. In all previous figures this remained constant (BF=.60, RS=.15, WS=.15, BS=.10). However, in Figures 8 and 9, these proportions of host species were changed to BF=.10, WS=.15, RS=.15, BS=.60, and BF=.15, WS=.15, RS=.60 BS=.10 respectively. Compared to Figure 4, Figure 12 indicates that by increasing the proportion of black spruce in 102 the stand, spring dispersal mortality is greatly increased. This is due to black spruce breaking bud late in the spring. Budworms dispersing to this host would tend to re-disperse. As a result, survival of re-dispers- ing larvae is reduced. By contrast, Figure 13 shows that increasing the proportion of red spruce, though not the most preferred host, does not markedly increase mortality in the spring. Implications in Forest Management The results of this study indicate that there is a relationship be- tween budworm survival during spring dispersal, the species composition and the open area of the surrounding stand. This, combined with the observations of Turner (1952) and Batzer (1969, 1976) of reduced damage to fir as the proportion of hardwoods increases, lends credibility to the adoption silvicultural techniques for management of spruce-fir in the northern boreal forest. First, by reducing the balsam fir in managed stands, we can effectively manage against overwintering spruce budworm survival. The reduction of preferred budworm overwintering sites (balsam fir staminate flower bracts) in a spruce—fir stand would facilitate this population reduction. By moving away from a pre-climax or climax forest, we may also manage against budworm. Analyses indicate that in the spring, and to some extent the fall, dispersal losses can indeed be influenced by canopy Openings and non-hosts. Therefore, management practices presently used in spruce-fir silvicultural systems, such as large clearcuttings and shelterwood methods, should be modified or replaced. Both clear- cuttings and shelterwood techniques perpetuate even-aged stands of pre- dominantly balsam fir (depending on site conditions). These stands are considered highly susceptable to budworm attack (Hatcher 1960). Selection 103 cuttings, though more difficult, encourage uneven stand structure and at the same time allow for regulation of reproduction, canopy openings, and species composition. Using these methods, large susceptable aggregations of balsam fir could be discouraged while canopy openings and white spruce are encouraged. Further, non-host hardwoods in the maximum pr0portions economically allowed could be encouraged to further retard budworm out- breaks during the spruce-fir rotation. The diversification of the forest with canopy openings and non-host species would also serve to increase predators and parasites of budworm by increasing predator and parasite alternate host habitats (Kemp and Simmons, in press). This would be a very important, though indirect effect, of interspersing hardwoods and open areas in spruce-fir stands. Probably not all stands are ameanable to improved silvicultural practices. Large expanses of spruce-fir in parts of Maine are due in part to site conditions that are not tolerated by hardwoods. However, in some parts of Maine strip clearcuttings are used commercially and show potential in management of the forest against the budworm. In areas where accessi- bility is facilitated by road networks, patch clearcut harvesting is now being conducted. This method of harvesting also shows potential as a forest management practice that minimizes budworm impact. Small areas less than 10,000 acres that are intensively managed, as well as areas where there exists adequate varieties of species, both show excellent potential for management. Improved silviculture likely cannot be viewed as a panacea, but certainly would contribute to forest management by providing additional methods that, when integrated with other control methods (i.e. chemical), would lead to improved spruce-fir/spruce budworm management. 104 Figure 1. Program driver; driving routine for spruce budworm dispersal model. Initialize poleation, enter values for; probabilities, host non-host, species composition and weather. Distribute eggs over host range Call fall dispersal Calculate overwintering mortality Calculate Glypta and Apanteles parasitism _ n—u—d- Call spring diSpersal Print J// ( Stop 105 106 Figure 2. Fall dispersal subroutine for spruce budworm dispersal model. 107 3 Calculate T. minutum parasitism Calcualte larvae available to disperse after parasitism Calculate stand density factor Y Calculate larvae dis- persing as a function of Y and cloud cover Calculate larvae impact- ing host non-host and open area Calculate diSpersal mortality Rodistrihute larvae on host species /,. ( {eturn 108 Figure 3. Spring dispersal subroutine for spruce budworm dispersal model. 109 Calculate budworm available to dispersE] 6, 7 as function of parasitism J r1 Cal) ”VHIUC days YES NO . I Calculate the preference for eacn host 8 species as a function of total D. days i Aggregate larvae to host and non-host [Calculate stand density factor Y Calculate the number of larvae diSpersing as a func- tion of Y, cloud cover and preference function Calculate larvae impacting host, non-host and open as a function of diSpersal probabilities .——-—— Calculate diSpersal mortality as a function of open and non-host Redistribute larvae on host Species I ( Continue I\% (starvation) Calculate starvation mortality a. ..- (:5 Rettrn ) 110 Figure 4. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm.(Weather = Optimal, Open Area = 10%, Dispersal = 13, Random spacing). 111 ooo.oo~ ooo.om hwozuzoz hzmumua cosmos cosmos b I Aflflm OZHmmm ooo.o~ 60.0N bo.cv éo.om éo.om ba.oou AlIIHlHON 1N3383d 112 Figure 5. Graph of the effects on non-host percentages on dispersal mortality of spring and fall budworm. Host and non-host aggregated. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 10%, Dispersal = 13, Clumped Spacing). 113 ooo.mmm connoo pmozIzoz pzmummm Daemom Daemoc DDOHON Addh UZHmmm éo.om éo.om bo.oo~ AlIIHIHUN 1N3383d 114 Figure 6. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Dispersal time decreased. Figure 4 superimposed in hatched lines. (Weather = Optimal, Open Area = 10%, Dispersal = 8, Random Spacing). 115 emoszoz Hzmomum cosmos P - ooo.oo— aoo.oo cachet u: AAflh OZHmmm éo.om 69.0w bo.ccu AlITUIHON 1N3383d 116 Figure 7. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Weather during dispersal contained 25% cloudy days. Figure 4 superimposed in hatched lines.(Weather = 25% Cloudy, Open Area = 10%, Dispersal = 13, Random Spacing). 117 hmozuzoz hzmommm Daemon ooo.mo~ cachet . OODHDN o 023% ...: 3v bo.om 60.90 60.09" AlIWUIHON 1N3383d 118 Figure 8. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Weather during dispersal contained 75% cloudy days. Figure 4 superimposed in hatched lines-(Weather = 75% Cloudy, Open Area = 10% Dispersal = 13, Random Spacing). 119 hwozuzoz pzmumma sconce coonot 000.00" cannon ...—“Adm oszam $2 60.01 bo.om bo.oo 60.no— AlIIUlHON 1N3383d 120 Figure 9. Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Open area increased to 20%. Figure 4 superimposed in hatched lines.(Weather = Optimal, Open Area = 20%, Dispersal = 13, Random Spacing). 121 hwozIzoz hzuummm 000. P00— . 000.. 00 . 000.. 00 . 000.. 0v . 000.. 0N . 0 0 .00. on 3E .. ..00. 0* :00. 00 oszam :00. 00 ...00. 00" 11113180” 1N3383d Figure 10. 122 Graph of the effects of non-host percentages on dispersal mortality of spring and fall budworm. Open area increased to 30%. Figure 4 superimposed in hatched lines.(Weather = Optimal, Open Area = 30%, Dispersal = 13, Random Spacing). 123 ewoxuzoz HzmomML coauom . . 000H0v . 000n0N 000.m0— 1 AA