ll" \ l WWW!WNHWIHKIWIW”WW 4:» Io): (DOD—i 'r'r' vs Illllllllllllllllllllllllllllllllllilll 3 1293 01066 4906 This is to certify that the thesis entitled AN EVALUATION OF THE IMPACT OF THE SARATOGA SPITTLEBUG, APHROPHORA SARATOGENSIS (FITCH), ON THE GROWTH OF RED PINE, PINUS RESINOSA AIT. presented by Robert Lewis Heyd has been accepted towards fulfillment of the requirements for Ph-D . degree in m x”? . 7 - l, ,. ('\ {LEW/7 # (K ’ C--/ 14"‘7g/ Major professor Date 1’"? ’/' “)4 “'79 0-7 639 AN EVALUATION OF THE IMPACT OF THE SARATOGA SPITTLEBUG. APHROPHORA SARATOGENSIS (FITCH), ON THE GROWTH OF RED PINE, PINUS RESINOSA AIT. By Robert Lewis Heyd A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Forestry 1978 ABSTRACT AN EVALUATION OF THE IMPACT OF THE SARATOGA SPITTLEBUG, APHROPHORA SARATOGENSIS (FITCH), ON THE GROWTH OF RED PINE, PINUS RESINOSA AIT. By Robert Lewis Heyd The Saratoga spittlebug causes growth loss, deformity and mortal- ity of young red pine in plantations in the Lake States. The biological impact (or damage) of the Saratoga spittlebug was measured to provide basic data to construct an ecological impact model and to more effec- tively understand the socio-economic aspect of impact. Several studies were initiated to correlate feeding pressure from spittlebug populations on resulting tree growth and form. When trees were caged and exposed from light to heavy feeding pressure, an upper-whorl concentration of feeding activity occurred. Flagging, too, was concentrated in the upper whorls of the tree. In another study l6 plots were used representing different ground cover associations, soils, tree heights and ages. From these the following regressions were derived: Reduction of growth = -.558 + .leX; Reduction of photosynthetic potential = -.5l8 + .ZZlX; and Reduction of branch vigor = -.442 + .l36X where X is the log of the insect exposure per branch. These regressions were related to a useable Robert Lewis Heyd late-instar nymphal survey and tree-unit evaluation, and the following formula was derived: Y = -8.l40 + 53.897X where Y = mean insect expo- sure per upper-whorl branch and X = nymphs per tree-unit. This formula provided a way to relate feeding pressure from spittlebug populations to growth loss and lost growth potential. Different degrees of growth loss were related to tree deformity and mortality to develop a predictive model. A value of less than 25 percent growth loss showed only reduced growth of the tree and a small loss of potential growth for the following two-three years. Percentages from 25 to 40, however, showed trees having scattered partial branch flagging and some deformity from light sweep and large lower-bole limbs. Percentages from 40 to 70 provided some trees with top kill, whole branch flagging and serious degrade from sweep, crook, multiple stems and large lower limbs. Growth loss of 70 percent or greater led to numerous trees top killed, heavy degrade and a few to many dead trees. An analysis of a 45-year-old red pine plantation with a history of spittlebug injury showed that sweep, crook, fork, large and numerous lower bole limbs, wood stain, and growth loss resulted from spittlebug feeding. An ecological model was developed which incorporated ground cover association, insect population size, tree height and density, tree growth response, form, interior degrade, and susceptability to snow damage, with modifications of the physical environment. Dedicated to Sarah Garnet Heyd David Michael Heyd ii ACKNOWLEDGMENTS The opportunity for this study was provided through the USDA Forest Service, North Central Experimental Station, East Lansing, Michigan. I extend my appreciation to the Michigan Department of Natural Resources, the Packaging Corporation of America, and the U.S. Forest Service for providing research sites for this study. My deepest appreciation is extended to Dr. Louis Wilson, who provided guidance in the research and especially in the writing of this dissertation. I thank Dr. Wayne Myers, who gave council on the statis- tical analysis used herein; and George Heaton, biological technician, who assisted me in collecting and summary of the large amount of data required; and my mother, Elva Heyd and wife, Mary-Lynn, who sewed many sleeve cages. The constructive criticisms and helpful suggestions pro- vided by my committee members, Dr. Gary Simmons, Dr. James Hanover, and Dr. Donald Dickmann, are appreciated. My thanks to all members of the Forestry Department, Michigan State University, for their perpetual willingness to help--with a special thanks to Mrs. Joan Perry. Finally, a special thanks to my wife, Mary-Lynn, for her perseverance, love and understanding throughout this period of graduate study. TABLE OF CONTENTS Page LIST OF TABLES .................... . ....... Vi LIST OF FIGURES ........................... Vii INTRODUCTION ............................ l Objective ........................... 1 Impact Defined ........................ 2 BRIEF HISTORY OF THE SARATOGA SPITTLEBUG .............. 4 INJURY TO RED PINE ......................... 6 The Feeding Scar ....................... 7 Damage Effects on the Tree .................. 7 WITHIN TREE DISTRIBUTION OF INJURY ................. l0 Distribution of Spittlebug Feeding .............. lO Methods and Materials ................... lO Results .......................... l2 The Distribution of Flagging ................. l5 Methods and Materials ................... l5 Results .......................... l5 Discussion .......................... l7 EVALUATION OF TREE GROWTH RESPONSES ................. 19 Methods and Materials ..................... 20 Plot Selection ....................... 20 Plot Treatments ...................... 22 Feeding Frequency Determination .............. 24 Shoot, Needle, and Bud Measurements ............ 25 Tree Vigor Evaluation ................... 27 Statistical Analyses .................... 27 Results ............................ 29 Shoot Growth Response ................... 30 iv Page Photosynthetic Potential .................. 34 Branch Vigor ........................ 39 Discussion .......................... 46 PREDICTION OF INSECT EXPOSURES PER BRANCH .............. 50 Methods and Results ...................... 50 Discussion .......................... 53 PREDICTION OF TREE DEFORMITY AND MORTALITY ............. 54 Methods and Results ...................... 54 EXAMINATION OF THE GROWTH AND FORM OF TREES WITH HISTORIES OF SPITTLEBUG ATTACK .......................... 57 Methods ............................ 57 Results ............................ 58 Discussion .......................... 6l DISCUSSION ............................. 65 Ecological Model ....................... 65 Predictive Model ....................... 7O Socio-Economic Considerations ................. 73 LITERATURE CITED .......................... 77 Table -l LIST OF TABLES Tree heights and ground cover for all spittlebug plots in Lake County, Michigan, 1974-l976 .............. Red pine plot locations in Lake County, Michigan, and spittlebug sleeve cage treatments for l974-1976 . { ..... Tree heights, ages, and vigor for all plots, l975 ....... Plot groups formed by combining similarly treated plots, and number and duration of spittlebug treatments ........ Plot group correlation statistics between log of insect-days per branch and per branch-unit and reduction of growth. . . Between plot group similarities and differences within plot group combinations ..................... Plot group correlation statistics between log of insect-days per branch and per branch—unit and reduction of needle dry weight ......................... Plot group correlation statistics between log of insect-days per branch and per branch-unit and reduction of bud length ........................... Combined plot group correlation statistics between the log of insect-days per branch and per branch-unit and reduction of bud length .................. vi Page 21 23 28 29 31 33 37 43 45 LIST OF FIGURES Figure l 10 Location of Saratoga spittlebug research plantations in Michigan ......................... Distribution of scars and centimeters of branch per whorl from the feeding distribution study ............ Scar density on (X) total branch area, (0) needle-bearing internodes less current growth, and GI) one and two year old main stem from the feeding distribution study ..... Distribution of spittlebug flagging by whorls in six study areas showing heavy to light injury: (X) trees flagged; (I) entire branch flagged; (0) partial branch flagged. . . . Log of insect-days per branch regressed on proportionate reduction of growth for the plot group combination PGl, P62, P64 and P67 ..................... Log of insect-days per branch-unit regressed on proportionate reduction of growth for the plot group combination PGl, P62, P64 and P67 ..................... Log of insect-days per branch regressed on proportionate reduction of needle dry weight for the plot group combination P61, P62, P64 and P67 ............. Log of insect-days per branch-unit regressed on proportionate reduction of needle dry weight for the plot group combination P61, P62, PG4 and P67 ............. Log of insect-days per branch regressed on proportionate reduction of bud length for plot group combination P61, P62, P65, P68 and P69 ................... Log of insect-days per branch-unit regressed on proportionate reduction of bud length for the plot group combination P61, P62, P65, P68 and P69 ................ vii Page ll l3 T4 16 35 36 40 41 47 48 Figure Page ll Average scars per centimeter of two-year-old branch inter- node per tree regressed on average insect-days per branch per tree ...................... 52 l2 Forty-five year old plantation with current spittlebug injury: (A) sweet-fern pocket showing current mortality; (B) extreme sweep and multiple stems of a pocket tree; (C) large lower limb characteristic of spittlebug injured trees; and (D) sweep from spittlebug injury and snow damage ........................... 59 T3 Logitudinal sections of red pine trees from a 45-year-old plantation showing residual spittlebug feeding injury: (A and B) crook due to lateral branch dominance after spittlebug injury; (C) necrosis and stain of inner growth rings and a knot; and (D) extreme necrosis extending from inner growth rings to outer surface of bole showing com- pensatory growth on the opposite side of the bole ..... 62 l4 Ecological model of the Saratoga spittlebug .......... 66 15 Population dynamics model of the Saratoga spittlebug--red pine ecosystem ....................... 69 T6 Predictive model for Saratoga spittlebug attack ........ 7l viii INTRODUCTION The Saratoga spittlebug, Aphrophora saratggensis (Fitch), an insect native to Eastern North America, is a serious pest of pine in the Lake States, especially in northern Wisconsin and Michigan, because it severely deforms and kills young planted pines. Red pine, Pinus resinosa Ait., is the preferred host, but jack pine, Pinus banksiana Lamb and other pines are also attacked to a lesser degree. Although easy to suppress, control decisions for the spittlebug are often based on speculation. Ewan (l96l) predicted a "severe damage" threshold given the number of insects per unit of tree size and density. This is useful, but knowledge of the effects of populations below severe damage threshold is necessary if the forest manager is to make respon- sible management decisions. The Saratoga spittlebug causes tree deform— ities such as limbiness, large knots, crook and sweep, and loss of wood volume from growth loss and/or tree mortality. Thus, control decisions should include criteria for evaluating deformity and growth loss thres- holds as well as other multiple use values. Objective The objective of this study was to determine mortality, deformity, and growth loss caused by Saratoga spittlebug feeding on young red pine in plantations. This provides the basic data to construct an ecological l impact model and to more effectively evaluate socio-economic impact. Impact Defined The impact of an insect population can be broadly defined as the cumulative net effects of the insect resulting in modification of management activities for specified forest resource uses and values (USDA 1972). Impact has two components: (1) ecological--the cumulative net effects of the insect on the total forest site and areas offsite; and (2) socio-economic--the value judgments and/or decision criteria established by management objectives (Averill, 1977; USDA l972). The ecological component of impact deals with the physical changes of the tree and the environment caused by insect activity. This component is as variable as the different types and degrees of insect injury and combinations of trees and environments. The ecological component pro- vides information needed to assess the socio-economic component. The socio-economic component is determined by management objectives as they differentially emphasize multiple use concepts (i.e. timber, wildlife, watershed and recreation). A positive, negative or neutral impact may result from an insect-tree interaction depending upon the management objective of the forest. Suppose timber production is the primary objective. An insect which randomly attacks an area of suppressed trees may kill many trees, but may also release the remaining trees producing an overall positive impact. Pocket-type infestations, however, may result in loss without significant release of trees and excessive limbiness in trees bordering the pockets and this may produce a negative impact. However, if wildlife management is the objective in this situation, a positive impact may result. Pockets create edge and potential forage areas, both of which enhance wildlife values. The interactions are numerous and complex. To evaluate the socio-economic component of impact the specific management objectives and values must be known. Ecological impact in this dissertation refers to the effects of spittlebug feeding injury upon growth rate, tree form, and mortality as modified by the site. These qualities affect the suitability of the tree as a forest product and a component of multiple use concepts. BRIEF LIFE HISTORY OF THE SARATOGA SPITTLEBUG The Saratoga spittlebug overwinters as an egg tucked under the bud scales of the larger buds in the upper whorls of the pine host. The nymphs eclose in early May, drop to the ground and search for and feed upon understory vegetation, primarily within ten feet of the tree. The five nymphal instars feed upon these alternate hosts in spittle- masses for about six weeks. Many herbacious and woody plants serve as suitable nymphal hosts. However, heavy infestations are usually cor- related with the density of sweet-fern, Comptonia peregrina (L.) Coult. If 35 percent or more of the ground cover is sweet-fern, or if it is present in certain combinations with other alternate hosts, moderate to heavy damage often results (Wilson, Heaton and Kennedy, 1977). Adult transformation generally occurs in late June peaking before mid-July. Newly emerged adults move to and feed upon a pine host. The needle bearing internodes of the tree represent the feeding universe (Ewan, l96l). Adults feed two to five times daily depending upon the time of feeding and the temperature. Peak feeding occurs two-three weeks following emergence, and each adult averages 3.5 feeding punctures per day. About 90 percent of the feeding damage is complete by mid- August. In September the few surviving adults feed about 1.5 times each day. The seasonal average is 2.63 feeding punctures per day (Ewan, l96l). 4 Oviposition starts one or two weeks after adult transformation and peaks in late July or early August. Each female averages l4.6 eggs (Ewan, l961). INJURY T0 RED PINE Spittlebugs commonly damage pines two to 15 feet tall. Larger trees may be attacked but a plantation with 6' X 6' spacing generally approaches hosts needed for nymphal survival. Also, as young red pines double in height, the number of branches at least triples (Miller, l965). Thus larger trees can withstand much more feeding because of the increased total branch area. The Saratoga spittlebug causes growth loss, chlorosis and necrosis of branches and branchlets (flagging), top kill, tree mortality, and deformity (limbiness, large knots, crook and sweep). When flagging first appears, the tree has already incurred irrepairable growth loss, and if flagging encompasses entire branches or whorls, deformity occurs. Top kill or severe upper-whorl damage causes a lateral branch to become the leader. Ewan (l961) reported that damage symptoms normally appear the year following feeding unless feeding is extremely heavy. Then, flagging can occur at the end of the same growing season. He also noted the time from first scattered flagging to entire tree mortality was one-two years. Spittlebug damage generally occurs in pockets corresponding to areas of suitable ground vegetation within the plantation. These pockets can enlarge and coalesce if there are adequate numbers of suitable alternate hosts. The Feeding Scar The adult spittlebug damages the tree by sucking sap and leaving a residual feeding scar which serves as a permanent block to xylem and phloem conduction. The first evidence of necrosis occurs within 24 to 48 hours after feeding (Anderson, 1947). This is followed by resin accumulation and spreading of the scar area. A disalignment of trachieds also occurs on the side opposite the transport block (Ewan, l96l). Ewan also found the carbohydrate content of spittlebug injured shoots and roots is lower than that of unattacked trees. This damage is characteristic of moisture stress but differs from girdling which accumulates carbohydrates distal to the girdled area. Ewan (l96l) noted that feeding scars were concentrated on the one-year-old internodes (last year's growth). Older needle-bearing internodes and the current growth had less. He also reported feeding scar densities did not differ between the top and bottom surfaces of branches, but the phloem had 30 percent more feeding scars than the xylem. He found no differences in the distribution of scars between branch whorls within the tree. Damage Effects on the Tree In the Lake States, adult Saratoga spittlebug populations peak in mid-July after red pine shoot elongation is nearly complete. Needle elongation has begun at this time and new buds approximately 40 to 50 percent developed (Sucoff, l97l). Thus, feeding injury does not affect current shoot growth, but it can influence needle elongation and bud development. '1' The developing buds and the current needles are largely respon- sible for auxin production. Auxin-directed translocation is responsible for the movement of nutrients from storage cells to young growing tissues (Kozlowski, 1964; Kulman, 1965). Water stress, then, induced by withdrawal of liquid from feeding, plus increases in xylem and phloem resistance from the residual feeding scar, inhibits the growth of the terminal meristems and new needle growth. This further decreases auxin production in spittlebug injured branches and reduces translocation to the damaged area, thus reducing needle growth and photosynthesis. Also, greater auxin production by terminal meristems is responsible for. inhibition of lateral buds (Kozlowski, T964; Kozlowski and Winget, l964). Thus, given less terminal bud development in upper whorls, greater growth of unattacked or less severely attacked lower whorl branches will result. From this, lower whorl branches compete for dominance with a net loss of unidirectional height growth. The following year, the one- year-old needles which are the most important source of food for shoot growth (Gordon and Larson, l970; Dickmann and Kozlowski, l968; Kozlowski and Winget, l964) will be shorter; plus, the expanding bud will have fewer leaf primordia due to the less than optimal conditions for forma- tion in the previous season. Adverse environmental conditions in addition to spittlebug attack would compound the stress. Even if spittlebug populations are controlled the following year, the shorter needles from the previous season and smaller buds will reduce photo- synthetic potential for at least three years. Any decrease in photo- synthetic potential affects the entire tree--needles, stems and roots. In addition, the feeding scars persist as a block to water flow in the xylem until stem diameter increases significantly. Aggregates of scars cause necrotic areas which may serve to block phloem and xylem transport for several years. From another viewpoint, decreased auxin supply and decreased carbohydrate result in decreased tracheid diameters which may increase xylary resistance and decrease wood production (Larson, 1962, 1963). A change in the growth pattern of spittlebug injured red pine was noted by Ewan (1961). He calculated the ratio of internode lengths within whorls for a given year to the terminal growth for that year, using the lateral/terminal ratio (L/T ratio) developed by Benjamin et a1., 1953. They established L/T ratio limits called zones of normalcy within which normal development occurs. A characteristic pattern of longer internodes in upper whorls progressed to shorter internodes in lower whorls. Ewan used the L/T ratio to indicate growth imbalance in attacked trees (i.e. loss of upper whorl dominance and/or asymmetry of crown). He reported that growth imbalance was evident, but the L/T ratio was an indicator only for long-duration light to moderate infesta- tions. The damage of rapidly developing heavy infestations could not be predicted due to the year delay of damage symptoms. WITHIN TREE DISTRIBUTION OF INJURY A knowledge of the distribution of spittlebug injury within a red pine is necessary in understanding tree growth response. The objectives of this portion of the study were to determine the distribu- tion of feeding and flagging within the tree, and to relate this to visible signs of tree injury. Distribution of Spittlebug Feeding_ Methods and Materials In the summer of 1975, six red pines about six feet tall in a 12-year-old plantation (T18N, R12W, Sec. 23, Lake County) (Figure 1) were enclosed in 6.5' X 6.5' X 6.5' cages made of 20 X 20 per inch mesh saran screening. Treatments consisted of 25, 50, 75, 100, 150 and 200 spittlebugs per cage for 21 days (first three weeks of August). Insects were allowed to die normally without replacement. Natural populations seldom exceed 100 insects on trees of this size. Only unattacked trees were selected so the insects would not be influenced by previous spittle- bug feeding. At the end of the test live insects were counted in each cage. The trees were then cut at the ground line, and the bark was carefully removed using a blunt knife so as not to damage the scars on the xylem surface. The length of all stem and branch internodes was measured and the number of scars were counted and recorded by location. 10 11 I24M ' I23M IZZM 12'" WEXFORD tzou. ‘F 1' tmw. o T. .r tmw. . Inn. LAKE OSCTOLA RJ4W. R.I3W. RJZW. RIIW. RIOW. R.9W. R.8W. R.7W. Figure 1.--Location of Saratoga spittlebug research plantations in Michigan. 12 Results The number of scars on each whorl generally paralleled branch size as measured by total centimeters of internode (Figure 2). This suggests a somewhat even distribution of feeding throughout the tree. Analysis of scar density (Figure 3), however, revealed more scars per branch length in upper whorls than in lower whorls in all treatments except the 25-insect exposure. In the latter, the low feeding pressure did not result in any areas of high scar density. The highest scar densities were found on the one and two-year- old main branch stems followed by all the needle bearing internodes except the current growth (Figure 3). There was a proportionately greater amount of one and two-year-old main branch stem available relative to less preferred sites in upper-whorl branches. This may account for the upper-whorl feeding preference. However, scar density on the little preferred non-needle-bearing internodes was also greater in upper whorls. In the ZOO—insect treatment, the bare internode scar density of the fourth and fifth whorls actually exceeded that of the total branch in lower whorls. This may be due to the greater moisture and nutrient content of upper-whorl branches (Mamaev, 1956; Clausen and Kozlowski, 1967; Szymanski and Szczerbinski, 1962). The spittlebug may not actively select upper whorls to land on an indicated by the general parallel between number of scars and branch size; however, upon landing in an upper whorl, the high nutrient and moisture conditions may cause prolonged and/or repeated feeding. The 50, 100 and 150 insect treatments showed more scars than expected per branch length on the one and two-year-old main branch stems NUMBER OF SCARSi') OR CENTIMETERS 0F BRANCH“) l3 200 ISOINSECTS IOOOd- INSECTS I 800“ ' 6007 4M30F’ 26M) IOOINSECTS TfilNSECTS )- l. l- i i i— i 25INSECT IOOO'" OCH}- 600‘ 20!) ’3. .- A- , 5’19” WHORL Figure 2.--Distribution of scars and centimeters of branch per whorl from the feeding distribution study. NUMBER OF SCARS PER CENTIMETER 0F BRANCH 14 200 INSECTS ISO INSECTS 4 100 INSECTS L n 1 fl I r l T 75 INSECTS T 1P e e O O 4. 5 50 INSECTS 25 INSECTS WHORL Figure 3.--Scar density on (X) total branch area, (0) needle-bearing internodes less current growth, and (I) one and two year old main stem from the feeding distribution study. 15 in lower-whorl branches (Figure 3). The branches were larger in these whorls compared to surrounding whorls, and probably received more than their share of insects. The needle bearing internodes did not show higher scar densities in lower whorls. This may be due to the lower proportion of one and two-year-old main branch stems in the needle bearing internodes of lower-whorl branches. The Distribution of Flagging Methods and Materials Seven plots within five red pine plantations with spittlebug ' flagged branches were examined in August, 1976. The locations of the plantations (Figure l) and plots in Michigan were: Plot Number Location County 1 T18N,RllW, Sec. 6, SW% Lake 2 T24N,R9W, Sec. 16, NE% Wexford 3 T18N,R9W, Sec. 4, NEk Osceola 4 T18N,R9W, Sec. 4, NEk Osceola 5 T19N,R12W, Sec. 3, NW% Lake 6 T19N,R12W, Sec. 3, NW% Lake 7 T18N,R9W, Sec. 3, NW% Osceola The trees ranged from 4.5 to 14.0 feet tall, and contained light to severe injury. A one tenth-acre plot was established in each planta- tion, and the distribution of flagged branches and the proportion of flagging were recorded for each whorl of each tree. Results Analysis of the first six plots showed flagging was highest in the upper whorls and diminished with declining whorl position (Figure 4). 16 09- 0.3.. 0-7- 0°6- 05‘ 0-4, 03' 0:2- 0- I - 0.9. 0.8-. 0.7.: 0'6" 0.5.. 0-4. 0-3, 0-2« 0-1, PROPORTION OF TOTAL 0-9.; n=58 .. n=73 O'B‘r “ 0.7+ .. 0-6« 1- 0.5..- ._ 0-4.. l O'3ar 4" 0.24. q.- 0.. .‘L ur- . L fine—.4— WHORL Figure 4.--Distribution of spittlebug flagging by WhOFIS in Six study areas showing heavy to li ht injury: (X) trees flagged; (I) entire branch flagged; (03 partial branch flagged. 17 The frequency of branches displaying partial flagging was greatest in the third whorl because second whorl branches primarily exhibit whole branch flagging. Second whorl branches were composed of only one internode of preferred feeding sites and a whorl of current growth. Generally, the one-year-old internode received a lethal scar density first, causing the entire branch to flag. Also, the mortality of the current whorl was due mainly to girdling of the one and two-year-old bole and not from feeding directly on the whorl. The seventh plot was more heavily flagged than the others because of two consecutive years of heavy feeding. All of the trees had some degree of flagging, and, only in this plot, the flagging extended lower than the fifth whorl. The first year of heavy feeding killed the top three or four whorls in most trees. The buds remained intact, however, allowing the adults to successfully overwinter. In the spring the lush ground cover with abundant sweet-fern supported a high population of spittlebugs, and resulted in heavy flagging of the remaining whorls. As in the other six plots, flagging severity diminished with lower whorl_ position, showing that the remaining higher whorls were still preferred even following top kill. Discussion The Saratoga spittlebug feeds primarily on the needle bearing internodes of red pine and preferably on the one and two-year-old main branch stem. Although feeding activity is distributed throughout the tree, greater scar density occurs in the upper whorls. Flagging also appears first in upper whorls. Thus, both studies support using 18 upper whorls to determine the influence of feeding injury upon tree growth. This information helps explain the increased number and size of limbs, fork, crook and sweep associated with spittlebug injured trees. By damaging upper whorls, the spittlebug causes tree deformity as well as growth loss and mortality. The upper whorls suppress the growth of lower whorls, maintaining the desirable conical shape of the tree (Pharis, 1976; Forward and Nolan, 1962). Thus, feeding imposed on the upper whorls causes lower whorls to compete for dominance, with a net loss of unidirectional height growth and an increase in girth. This may result in limbiness and large knots in the lower bole, fork and/or crook if upper whorls are stressed enough to allow lower-whorl branches to assume dominance. EVALUATION OF TREE GROWTH RESPONSES An evaluation of growth loss of red pine from Saratoga spittle- bug attack is an essential step in assessing impact. The observed parameters should reflect the growth responses of the tree accurately and meaningfully. It is well known that the length of a bud in red pine is directly related and highly correlated to the length of the resulting shoot (Kozlowski, et a1., 1972; Hanover, 1963; Szymanski and Szczerbinski, 1962). There is also a high correlation between bud length and the number of needle primordia which determine the number of fascicles on the next year's shoot (Kozlowski, et a1., 1972; Marion, et a1., 1968; Duff and Nolan, 1953). Thus, bud length reflects potential shoot length and photosynthetic potential, and therefore shoot vigor. The most vigorous buds (largest) are located in the upper whorls and they more accurately reflect the vigor of the tree. Also, bud length is a better indicator of potential shoot length for upper-whorl branches than for lower whorl branches or the terminal leader (Szymanski and Szczerbinski, 1962). These relationships between bud length and resulting shoot length and number of needle primordia suggest a method of evaluating the effects of different levels of feeding intensity upon growth. By measuring bud lengths on treated and control branches in upper whorls 19 20 of red pine and then measuring and comparing the resulting shoot lengths and needle dry weights the following year, growth response can be evaluated. Also, a ratio of bud length to the resulting shoot bud length will reflect changing shoot vigor. Methods and Materials The effect of feeding injury upon tree growth was examined by placing adult spittlebugs on red pine branches in sleeve cages. The effects of different intensities of feeding injury were monitored by examining the following parameters: (1) the ratio of bud length and resulting shoot length as a measure of expressed growth potential, (2) the ratio of needle dry weight to bud length as a measure of expressed photosynthetic potential, and (3) the difference between bud length and the resulting shoot bud length as a measure of changing shoot vigor. Needle chlorosis and flagging were used as indicators of branch stress and mortality. Plot Selection Sixteen plots were selected in Lake County, Michigan, to repre- sent a variety of ground covers, tree heights and other conditions which characterized Saratoga spittlebug infested plantations throughout the Lake States (Table 1). Individual plot sites, however, were chosen to display uniformity of ground cover, little slope, and no sign of insect injury or presence of potentially damaging insect populations. Each plot was selected by delimiting an area inea stand with about 30 trees. Five trees were picked randomly in the plot for .mm=_a> omen“ m>wgmu ou ummmgm>m mm: uopn mgwpcm mg» No new Amucv mag» ammo 59mm merccaogczm Lm>ou ecsoguo 21 NN ep -1 me 11 - m mp m.m N.N 11 mp Np mN 11 m 11 11 11 mm N.N— m.m 11 m. N 11 11 11 11 11 m mm n.0, ¢.m 11 SN N N 11 11 1- N N— mN N.m N.N 11 mp m -1 11 -1 11 oN 11 mm N.m e.m 11 NF NP m m N 11 11 11 NN m.m N.N 11 FF o N o o_ 1- 11 N ¢N N.N o.N 11 op N 11 m m 11 11 11 em m.m o.N 11 a v m m m 11 11 11 cm m.m m.m 11 m o 11 11 11 1- 11 mm mm N.NF o.o_ ¢.w N N 11 11 11 11 o o_ Nm m.¢_ m.mp N.NP m m 11 11 11 N N om N _.¢_ N.NN m.op m 11 N N m mp op NN Nm m.m m.w m.m e e N m N N on NN oN N.N N.m N.m m o 11 11 11 11 11 11 em o.pp o.m N.N N op 11 11 -1 11 m N ow m.- N.op o.m N mmoz a mewduo -ewwwom 1wmmum 1mwnmu wwwmw maaam umummw mmaeu on_ mNaP «No, zansaz Nomaev agave: uo_a «Nucmusmav Lm>ou ecsogo mmLN mmmgo>< .mNmN1eNmN .cmmmzovz .Nucaou mxmA cw muopa usampuuwam ppm Lo» gm>ou ucsoem new mugmww; mosh-1., «Naop 22 experimentation. On each tree, two or three similar branches from the same whorl were used for testing. One branch was exposed to different insect feeding pressures. The other branches were controls. Upper whorl (third and fourth node) branches were used. The one-year- old branch whorl (second node) was not suitable because of the small area of preferred feeding site. Upper whorls were selected because spittlebugs prefer upper-whorls, because upper-whorl branches flag first, and because buds are larger and better represent shoot growth and vigor. Plot Treatments Sixteen plots were established in 1974 or 1975, and used for one, two, or three consecutive years. Sleeve cages were used to contain spittlebugs on the host branches for all feeding tests. Each cylindrical sleeve cage was 3' long by l' in diameter. The body of the cage was plastic window screen and the ends were muslin cloth. One end was tied around the branch base, the other end was tied shut beyond the branch tip and supported with a string to the branch above. When used, one of the control branches on each tree was also caged but was without insects. The other control branch was left uncaged. Cages were removed at the end of each year and placed on the tree again the next season if the procedure was repeated. Plot 1 contained 20 spittlebugs per test cage for 72 hours to evaluate within plot variation, and plot 2 included no spittlebugs to determine the effect of the cage alone on shoot development (Table 2). Plots 3 to 7 received 5, 10, 20, 40 and 80 spittlebugs per cage on 23 Table 2.--Red pine plot locations in Lake County, Michigan, and spittlebug sleeve cage treatments for 1974-1976. Spittlebugs per Spittlebug N:;g:r Location Years Treated test cage no.: Exposure 1. R. s. 1 2 3 4 5 ‘me 1 ZON 13w 11 1974,1975,1975 20 20 20 20 20 72 hrs. 2 2ON 13w 11 1974,1975,1975 o o o o o 72 hrs.a 3 l9N 12w 3 1974 5 1o 20 4o 80 72 hrs. 4 19N 12w 3 1974 5 1o 20 4o 80 72 hrs. 5 20M 13w 11 1974,1975,1975 5 10 2o 40 BO 72 hrs. 5 20M 13w 11 l974,l975,l976 5 1o 20 4o 80 72 hrs. 7 20N 13w 11 1974,1975,1975 5 1o 20 40 so 72 hrs. 8 19N 12w 27 1975 10 20 3o 40 50 72 hrs. 9 19N 12w 27 1975 10 20 30 4o 50 72 hrs. 10 19N 12w 27 1975-1975 10 20 30 4o 50 72 hrs. 11 19N 12w 27 1975-1975 10 20 30 4o 50 72 hrs. 12 18N 12w 23 1975-1975 10 20 3o 40 50 72 hrs. 13 18N 12w 23 1975-1975 10 20 30 4o 50 72 hrs. 14 ZON 13w 11 1975-1975 5 1o 15 20 25 30 days 15 19N 12w 27 1975-1975 5 1o 15 20 25 30 days 15 l9N 12w 27 1975 5 1o 15 20 25 30 days aEmpty cages placed on tree for this period. fourth whorl branches for 72 hours. represent a wide range of attacks. These numbers were selected to After the first year (1974) new treatments (plots 8 to 13) were made with 10, 20, 30, 40 and 60 spittle- bugs per cage on third whorl branches because the range was too broad in the 1974 tests and the feeding distribution study indicated an upper whorl feeding preference. Also, the new tests gave data that filled in 24 the intermediate ranges of insect feeding exposures. Plots 14 to 16 included 5, 10, 15, 20 and 25 spittlebugs respectively per cage for a 30 day period to simulate effects of long term, low intensity (chronic) feeding (Table 2). The 30rday exposure closely corresponded to peak feeding activity of the insect. Three test branches per whorl were used in plots 14 to 16 instead of two as in the other plots. One branch had a test cage with insects, another one had only a test cage, and the third was uncaged. This permitted assessment of the effects of the cage on growth response. The sleeve cages used here were modified slightly. The outer end was covered with a screen cone instead of a muslin sleeve. This allowed greater light penetration and better air circulation. The long exposure (30 days) necessitated checking the insect-occupied cages for dead insects twice weekly. Dead insects were removed and replaced with live ones caught in nearby plantations. Cages on branches without insects were also opened at the same intervals to equalize the test conditions. Most tests were repeated for one or two years on the same branches to simulate repeated infestations. Three plots (8, 9, 16) were tested for only one year to determine branch recovery following a simulated population decline or control program. Feedinngrequency Determination Living insects were counted at the end of each test and multi- plied by the number of days the test was run. When insects died they were considered dead for one-half the test period. In the 30-day expo- sures this was calculated from one recharge (i.e. twice a week 25 replacement of dead insects) to the next and accumulated. This pro- vided an index of insect-days (ID). That is, ID = j_x g_where j_is the number of insects alive for each day of the test and d.is the number of days of the test. At the end of the tests scars were counted on all branches after removing the bark. The number of scars per branch was divided by the insect-days to determine the mean number of scars per insect-day or the feeding frequency (F). That is, F = ZS/ID where BS is the sum of scars. Branch size, or the amount of branch available to the exposed insects, was estimated to adjust insect-days to an equal volume of feeding material. A branch-unit index (BUI) was derived using the sum of all the bud lengths on the test branches. That is, BUI = b + b l 2 . bn where b]...bn are bud lengths in millimeters. The number and size of buds were representative of branch size and vigor, and after treatment, equal shoot lengths were not equal in vigor due to different treatment. Insect-days per branch-unit was determined by ID/BUI. Shoot, Needle, and Bud Measurements A growth factor based on the ratio of shoot length to bud length was used to determine growth loss following spittlebug feeding. On each branch, terminal and terminal-lateral bud lengths of the leader shoot and terminals of all lateral shoots were measured. Lengths of the shoots resulting from these buds were measured in September when fully developed. The mean bud and shoot lengths per branch were used to calculate growth factors. 26 The difference in growth between treated and control branches was determined by subtracting the growth factor of the treated branch (GFt) from the growth factor of the control branch (GFC) and dividing by the growth factor of the control. That is, the growth difference (GD) was: 6D = (6FC - GFt)/GFC. The control branch growth factor represented the normal or expected growth factor. Similarly, the effects of insect exposures upon the ratio of needle dry weight to bud length was used to determine the reduction of photosynthetic potential. All needles were collected after tests were terminated in September, 1976. The needles, separated by year, were carefully removed and placed in small paper bags. They were dried at 75 C for three days (until loss of weight reached an equilibrium) and weighed to the nearest 0.0l gram. Needle dry weights were then divided by the total length of buds which produced the needles. (Bud length is highly correlated to the number of needle primordia.) The difference in needle dry weight per unit bud length (NOW) was determined as follows: NOW = (NOWC - NDWt)/NDWC, where NOWc = control branch value, and NDWt = treated branch value. The effects of the insect exposures upon the change in average bud length per branch each year was also observed to quantify changes in branch vigor. On each branch, terminal and terminal-lateral bud lengths of the leader shoot and terminals of all lateral shoots were measured. The mean bud lengths per branch were used for the comparisons. Bud lengths were measured before placement of the sleeve cages in the first season of treatment and remeasured each year thereafter in 27 September and early April before bud swell to test for differences between these periods. The difference in bud length (BLD) between the control and experimental branch was calculated by subtracting the treated branch difference in average bud length (BLDt) from one measurement period to the next from the control branch difference in average bud length (BLDG) and dividing by the control branch average difference as an estimate of the normal or expected difference. That is, BLD = (BLDc - BLDt)/BLDC. Tree Vigor Evaluation A means of determining plot differences in tree vigor was needed to help understand possible differences in treatment response. The resulting measure of vigor must take into account that branches were treated, not the entire tree. The size, or vigor, of a branch within a whorl differs between trees. Thus, a measure of tree vigor and branch vigor was combined to produce a tree vigor index. The mean height of the trees within a plot was divided by the tree age (number of whorls) to give a relative measure of site quality (or tree vigor). This quantity was then multiplied by the mean total bud length per branch per plot as a measure of branch vigor. This gave an overall vigor index for each plot (Table 3). Statistical Analyses Data from similarly treated plots were combined into groups for statistical analysis (Table 4). The results were treated using an 28 Table 3.--Tree heights, ages, and vigor for all plots, 1975. "mgr-14316211.)- T13: . "52:41:11.5: 1:33;: (feet (years) :SE(mm)b 1 10.7 12 89 i 8 79 2 9.6 12 102 i 8 81 3 6.7 12 84 i 8 47 4 8.5 12 93 i 9 66 5 12.8 13 98 i 7 97 6 13.3 13 85 i 7 87 7 10.0 13 73 i 7 58 8 8.6 9 170 i 17 162 9 7.0 9 105 i 8 81 10 7.0 9 96 i 9 74 11 8.1 9 138 i 10 124 12 8.4 11 122 i 9 93 13 7.8 11 115 i 5 81 14 9.4 12 140 i 14 109 15 9.8 9 116 i 14 126 16 8.2 9 101 i 11 92 aAge = no. of whorls bStandard error (n = 10) cVigor index = (tree height/age)(total bud length/branch) analysis of covariance with dummy variables. Insect-days per branch- unit and insect-days per branch each were regressed on proportionate differences in growth, bud length, and needle dry weights as determined by the treatment vs. control branch paired comparisons. Insect-days per branch were regressed to test the significance of the treatment- 29 Table 4.--Plot groups formed by combining similarly treated plots, and number and duration of spittlebug treatments. Number of Duration of 2:33p Cogggfizd Consecutive Treatment Treatments (days) P61 1,3,4,5,6,7 l 3 P62 l,5,6,7a 2 3 P63 1,5,6,7 3 3 P64 8,9,10,11,12,l3 1 3 P65 10,11,12,13 2 3 P66 8,9 l 3 P67 14,15,16 1 30 P68 14,15 2 30 P69 16 1 3O aPlots 3 and 4 discontinued after first year due to flare-up of endemic spittlebug population. response correlations without considering the amount of branch area exposed or the branch vigor. The analysis of covariance determined if a significant relationship existed between insect exposure per branch and the observed parameters, and if there were between plot differences in treatment response. Results To more effectively interpret the results of these studies, the influence of the acute and chronic sleeve cage environments upon the observed growth responses was evaluated. A paired t-test was used to examine differences in response between the caged and uncaged control branches of the chronic plots (plots 14,15,16) and of the control plot 30 (plot 1). No significant differences were found; thus, treatment response was not influenced by either sleeve cage environment. Also, tree vigor may significantly affect the response of red pine to Saratoga spittlebug attack. Thus, to help broaden our under- standing of factors influencing tree vigor, the soil and ground cover from each site was analyzed and correlated with tree vigor measured as mean tree height/plot age. The percent area occupied by each plant species and bare ground on the floor of each plot was determined (see Table 1), and a soil pit was dug in each plot to determine the depth, thickness, texture (Bouyoucos hydrometer method) and particle sizes of the sand fraction for each horizon and color band. No significant cor- relations were found between any of the soil measurements and mean tree height/plot age, or between percent abundance of any plant species or bare ground and tree height per plot age, or any combination thereof. Thus, soil and ground cover analysis was not needed to interpret the effects of site on tree vigor, and subsequently on growth responses. Shoot Growth Response When growth differences were regressed on the log of insect-days per branch and per branch-unit for plot groups 1, 2, 4 and 7 (Table 5), the branch-unit values gave higher correlations (r-values). Plot group 4 gave the lowest correlation because it had the largest branches and the narrowest range of insect exposures. This diluted insect exposures consequently increasing response variability. Plot groups 1 and 4 each had a deviant plot which was eliminated in the regressions (i.e. plots 3 and 8 were significantly different from 31 Table 5.--P10t group correlation statistics between log of insect-days per branch and per branch-unit and reduction of growth.a Igzegt- Insect-days Igzegt- Insect-days Plot e: per Plot Group e{ per Group Baanch Branch-unit Combination Baanch Branch-unit P61 .89b .93 951, P62 .85 .90 .14c .12 .15 .13 n=30 n=30 n=43 d n=43 d (N.S.) (N.S.) P62 .82 .88 P64, P67 .80 .88 .17 .14 .17 .13 n=18 n=l8 n=4O n=40 (N.S.) (N.S.) P64 .69 .75 P61,P64,P67 .77 .86 .10 .09 .18 .14 n=30 n=30 n=65 n=65 (.84) (.89) P67 .76 .88 P61, P62 .78 .86 .20 .14 P64, P67 .18 .14 n=15 n=15 n=83 n=83 (.84) (.89) aCombinations exclude deviant plots. 0 CStandard error of the estimate. d Correlation coefficient (all significant at the 1 percent level of probability). If plot groups are significantly different, the correlation coefficient taking this difference into consideration is given to show improvement in correlation if plot group differences could be accounted for quantitatively; if not given, N.S. = not significant. their respective plot groups). Plot 3 in P61 was the least vigorous of all plots (see Table 3) and displayed 23 percent greater growth loss; however, a portion of this loss may have been due to unrecorded feeding from an endemic spittlebug population. Plot 8 in P64 was by far the most 32 vigorous of all plots and displayed 12 percent less growth reduction. Thus, growth loss may be over- or under-estimated by the given percent- ages with trees of very low or very high vigor. Plot groups were further combined to test differences between plot groups (Table 5). That is, comparisons were made between years, consecutive treatments (accumulated), whorls, chronic and acute expo- sures, and plot locations. No significant difference was found between P61 and P62 (Table 5). This supports accumulating insect exposures for two years and indicates no significant difference in response between years (Table 6). Thus, the feeding pressure of the previous year should be added to that of the current year when assessing growth loss. No significant difference was found between P64 and P67, indi- cating the acute and chronic treatments and the different locations of the two plot groups elicited similar responses. Thus, the regression applies to growth loss incurred by different types of feeding pressure at different locations. The plot group combinations 1, 4 and 7, and l, 2, 4, and 7 included significant differences which may be due to differences in whorl position, location of plot, tree vigor (Table 6) or some combina- tion of these and others. However, without considering group differences the correlations by themselves were significant at the 1 percent level (Table 5). Reasonable correlations occurred when all plot groups (excluding plots 3 and 8) were used. Both insect exposure per branch and insect 33 Table 6.--Between plot group similarities and differences within plot group combinations. Plot Group Combinations Similarities Differences PG1, PG2 PG4, PG7 P61,PG4,P67 P52,P54,P57a P52,P55,P58b PG1,PGZ PG4,PG7 PG1,PGZ P65 P58,P596 Same branches Same treatments Whorl treated Year of treatment First year of treatment Year of treatment Second year of treatment Years of treatment P62--tw0 years of treatment Acute vs. chronic exposures Plot locations and conditions Year of treatment Acute vs. chronic exposures Whorl treated Plot locations and conditions P62--two years of treatment Acute vs. chronic exposures Whorl treated Plot locations and conditions Acute vs. chronic exposures Whorl treated Year of treatment Plot locations and conditions P62--two years of treatment Year of treatment Acute vs. chronic exposures Whorl treated Plot locations and conditions P62 and P68--tw0 years of treatment P69--0ne year of treatment, then one year without Year of treatment Acute vs. chronic exposures Whorl treated Plot locations and conditions aUsed in study of photosynthetic potential. bUsed in study of branch vigor. 34 exposure per brancheunit related directly to proportionate growth reduction (Figures 5 and 6). Thus, a prediction of growth loss given insect-days per branch or per branch-unit without considering the between plot group differences is a good, broad-based predictive tool. However, growth loss may be over- or underestimated with trees of very low or very high vigor. Photosynthetic Potential When differences in needle dry weight (NOW) were regressed on insect-days per branch and per branch-unit for plot groups 1, 2, 4 and 7 (Table 7), the branch-unit values gave stronger correlations than the branch values alone. Plot group 4 displayed the weakest correlations as in the growth response study. Plot groups 1 and 4 each had a deviant plot which was eliminated in the regressions (i.e. plots 1 and 8 were significantly different from their respective plot groups). Plot 1 in PGl consisted of five equal low-intensity insect exposures, and displayed significantly less NOW reduction. The regression line in plot 1 was not representative of a range of exposures, and, therefore, a comparison with other plots was invalid. It was included in the regression to provide additional measurement of low-intensity treatment-response variability. Plot 8 in PG4 was the most vigorous of all plots in all plot groups and displayed less reduction of NDW in the insect—days per branch analysis only. Unlike the growth response study, the branch-unit modification accounted for the greater vigor of plot 8. Thus, unless 35 LOO 2 EFF-F?— 7 r = .7"7€5 n = 83 2 .804 Y = -.558 + .211x 2 .60d .403 PROPORTIONATE REDUCTION OF GROWTH "' o 20 v r v 1 4f 3 fl I fl l ' 2.00 3.00 4.00 5.00 6.00 7100 8.00 LOG OF INSECT-DAYS PER BRANCH Figure 5.--Log of insect-days per branch regressed on proportionate reduction of growth for the plot group combination P61, P62, P64 and P67. 36 1.00 *2 T 7 r = .ENSC) n = 83 2 I 80 Y= .4l0+.234X 2 1— ' ' 3 O (r (9 LL. .GK)‘ 0 Z 9 B D .40. 0 LL! c: LL! '3 .20- Z 9 . l- 8 Q .00q O 0: 0L -0 O ' l 1 v f T ' I ‘ I ' -3.00 -2.00 -I.00 0 1.00 2.00 3.00 LOG 0F INSECT-DAYS PER BRANCH-UNIT Figure 6.--Log of insect-days per branch-unit regressed on proportionate reduction of growth for the plot group combination P61, P62, PG4 and P67. 37 Table 7.--P10t group correlation statistics between log of insect-days per branch and per branch-unit and reduction of needle dry weight.a Plot 183;:t‘ Insegt-days Plot Group Igzsgt- Insect-days Group per er . Combination per per . Branch Branch-un1t Branch Branch un1t P51 .82: .85 P61 , P52 .71 .74 .14 .12 .22 .22 11=2od n=20 n=33 n=33 (.79)e (.85) P02 .80 .85 PG4, PG7 .85 .90 .17 .17 .15 .12 n=18 n=18 n=40 n=40 (.87) (N.S.) PG4 .65 .67 PGl,PG4,PG7 .78 .84 .14 .14 .18 .16 n=30 n=30 n=55 n=60 (.84) (.88) PG7 .82 .88 PGZ,PG4,PG7 .84 .88 .12 .10 .16 .14 n=15 n=15 n=53 n=58 (N.S.) (N.S.) PG1,PGZ .78 .83 PG4,PG7 .19 .17 n=73 n=78 (.83) (.88) aCombinations exclude deviant plots. bCorrelation coefficient (all significant at the 1 percent level of probability). cStandard error of the estimate. dPlots 3 and 4 were discontinued after the first year due to a flare-up of an endemic spittlebug population. eIf plot groups are significantly different (5 percent level of probability). the correlation coefficient taking this difference into consideration is given to show improvement in correlation if plot group differences could be accounted for quantitatively; if not given, N.S. = not significant. 38 host vigor is evaluated, the reduction of photosynthetic potential (NOW) will be over-estimated with trees of very high vigor. Several plots in plot groups 1, 2 and 4 approached a significant difference at the 10 percent level of probability, which may be because bud length is highly correlated to number of needle primordia and not needle dry weight directly. Environmental factors influencing the growth of the needle primordia influence the expression of potential NOW and, therefore, increases response variability between locations and from one year to the next. Plot groups were combined to test between plot group differences in treatment response (Table 7). Plot groups 1 and 2 are significantly different indicating that one year and two consecutive years accumulated exposures do not relate and/or response is different between years of treatment. This is contrary to the findings of the growth study which showed no significant difference between P61 and P62. A significant difference was found between P64 and P67 in the insect-days per branch regression only. Perhaps the branch-unit analysis found no significant difference because branch size and vigor was con- sidered, thereby decreasing response variability. This indicated no actual difference between P64 and P67. Thus, the acute and chronic treatments, and the different locations elicited similar responses as in the growth response study. This was supported by the lack of a signifi- cant difference when combining P62, P64 and P67. This also indicated no difference between response of the different whorls, and that the two years accumulated exposures of P62 relate to one year's exposure of P64 39 and P67. Thus, the difference between P61 and P62 was primarily due to the different years of treatment, and the primary difference between P61, P62, P64 and P67 was the difference in response between years. Although a significant difference was not found between P62, P64 and P67, a significant difference was found between P61, P64 and P67. Again, this indicates that the between year of treatment difference was the primary factor in the plot group response differences of P61, P62, P64 and P67. By combining all plot groups (i.e. P61, P62, P64 and P67) a regression equation, which accounts for all plot variation, was derived to predict reduction of photosynthetic potential (i.e. needle dry weight) from insect-days per upper-whorl branch (Figures 7 and 8). The resulting correlation serves as a broad-based predictive tool. However, by using it, a reduction of photosynthetic potential may be overestimated if trees are highly vigorous as observed in plot 8. Branch Vigor Plot groups 1, 4 and 7 are summer (pretreatment) to Spring bud length comparisons after one treatment. Plot group 2 is a spring to spring comparison after two treatments and plot groups 5, 6, 8 and 9 are spring to fall comparisons after two treatments. Plot group 3 is a spring to fall comparison after three consecutive years of treatment. No significant difference was found between the fall and the following spring's average bud lengths. Thus, a treatment response is reflected equally at both times. 40 LOO .L H T = .777 7 n = 73 ' Y = '.518 + .ZZIX .80‘ .601 .40- .2£)< .00- 1 -.20 . L . , r PROPORTIONATE REDUCTION OF NEEDLE DRY WEIGHT 2.00 5.00 4.00 5.00 5.00 ' 7.00 ' 8.00 LOG 0F INSECT-DAYS PER BRANCH Figure 7.--Log of insect-days per branch regressed on proportionate reduction of needle dry weight for the plot group combination P61, P62, P64 and P67. 41 l— I.OO .5 :1: 192 r == 13213 7 0.1 n = 73 ' 7 3 Y = .483 + .243x >_ .801 a: o 4 21 = . 4 8 .501 7 4 7 LLI I 42 z 1 2 4 4 ES ‘1 1. 4 2 z .40‘ I ‘ 9 .- ._ II 4 8 | 4 2 4752 O I a 4 33 .204 ' 4 43, 111.1 -. 4.. ‘2‘ . o .: .004 a: o g 4 o: 1 ‘ 9- -.20 J? . —3.oo' -2:oo' -1.0o ' o ' 1.00 ' 2.00 ' 3.00 LOG 0F INSECT-DAYS PER BRANCH-UNIT Figure 8.-—Log of insect-days per branch-unit regressed on proportionate reduction of needle dry weight for the plot group combination P61, P62, P64 and P67. 42 / Differences in bud length were regressed on the log of insect- days per branch and per branch-unit for all plot groups (Table 8). The branch-unit values gave higher correlations than the branch values alone as in both the growth response and photosynthetic potential studies. The bud length correlations were generally weaker than the growth and photosynthetic potential analyses (Tables 6, 7 and 8), and as in the latter study several plots in P61, P62 and P64 approach sig- nificant difference at the 10 percent level of probability. This may be due in part to small differences in bud length (millimeters) being compared with a measurement error of approximately 0.5 millimeter and/or to the year to year sequence of environmental influences unique to each location which change the progression of bud lengths independently of treatment effects. No significant correlations were found in P64 and P66. These plots displayed the weakest correlations in both the growth response and photosynthetic studies because of their high vigor and low to moderate treatments which produced greater response variability. This along with the small differences in bud lengths and measurement error may explain the non-significant correlations. Plot group 6 consisted of two plots from PG4 which were not treated the second year; thus, a significant relationship would not be expected. However, P65 displayed a significant correlation in response to two consecutive seasons of treatment. Apparently, the accumulated exposures elicited large enough responses to overcome measurement error and response variability. Plot 43 Table 8.--Plot group correlation statistics between log of insect-days per branch and per branch-unit and reduction of bud length. Insect-days per Branch unit Insect-days per Branch unit Plot Insect-days Group per Branch Plot Insect—days Group per Branch P61 .583 .71 P55 N.S. N.S. .1o .10 n=30 n=30 P52 .77 .86 P57 .47d .59 .2o .15 .11 .09 n=18 n=18 n=15 n=15 P53 .73 .78 P58 .78 .86 .15 .14 .12 .1o n=l7 n=l7 n=10 n=10 P54 N.S.c N.S. P59 .93 .96 .12 .09 n=5 =5 P65 .75 .81 .05 .04 n=20 n=20 aCorrelation coefficient (all significant at the 1 percent level of probability). bStandard error of the estimate. cCorrelation not significant. dSignificant at the 7.7 percent level of probability--listed for comparative purposes. 10 in P65 displayed a greater reduction of branch vigor in response to treatment. This may be due to the low relative vigor of this plot in P65. This plot, however, was included when combining P65 with the other less vigorous plot groups. 44 Plot group 7 demonstrated a significant correlation between insect-days per branch-unit only. Again, this indicated the advantage of considering branch size and vigor. The weaker correlations of P67 in comparison to P68 and P69 were probably due to our inability to) interpret bud mortality in the spring and, therefore, including measure- ments of dead buds. Heavy flagging and branch mortality was found the following summer, but the chlorosis which developed in the spring was not interpreted as mortality. So, the buds were recorded as 20 to 40 percent bud length reductions instead of 100 percent loss. By the following summer, the results of P68 and P69 could be interpreted in view of the obvious flagging. Repetition of these severe treatments in P68 was less influencial because the already greatly stressed branches could lose little more. Thus, although P69 represented a year of treat- ment followed by a year of release, the correlation was strongest because of the severity of the treatments and little recovery. Plot groups were combined to test differences between locations, acute vs. chronic treatments, whorl position and years of treatment (Table 9). Significant differences were found with all combinations of plot groups. This may be due to the greater response variability of this study already explained. Also, unlike the growth response and photosynthetic potential analyses the average bud length per branch from one measurement period to the next has no relationship other than larger buds will produce longer shoots with larger buds on them and smaller buds will produce smaller shoots with smaller buds. The good correlation achieved when combining P61 and P62 (Table 9) lends support for accumulating insect exposures for two years. 45 Table 9.--Combined plot group correlation statistics between the log of insect-days per branch and per branch-unit and reduction of bud length.a Insects Insect- Plot days Inseg:;days Plot days Inse;:;days Groups Baggch Branch unit Groups Baggch Branch unit P61,P62 .73b .75 P61,PG7 .45 .55 .17c .15 .11 .11 n=48 n=48 n=45 n=45 PG1,PGZ .60 .69 PG2,PGS .59 .70 P63 .20 .18 P68 .19 .17 n=60 n=60 n=48 n=48 PG8,P69 .80 .84 PG1,P62 .69 .75 .13 .12 PGS,PGB .17 .15 n=30 n=30 P69 n=78 n=78 PG7,PGB .70 .76 P69 .16 .15 n=30 n=30 aCombinations exclude deviant plots; all combinations include significant differences. bCorrelation coefficient (all significant at the 1 percent level of probability). ‘ cStandard error of the estimate. However, the addition of P63 greatly weakened the correlation indicating a shortcoming in combining three consecutive accumulated treatments. This may be due to a small effect of two-year-old scars on branch vigor. Or perhaps, small losses of already stressed branches are hidden in measurement error and/or difference in response between years. Thus, two years of exposures may be the limit in correlating treatment with reduction of branch vigor. 46 Further examinations of the data suggested that P61, P62, P65, P68 and P69 could be combined to formulate the best prediction equations (Figures 9 and 10). These equations can be used to predict the effect of one or two years accumulated insect exposures on the reduction of branch vigor. Again, the branch-unit analysis produced a better cor- relation. Discussion Strong correlations were found between the insect exposures and reductions of shoot growth, photosynthetic potential and branch vigor. The branch-unit analyses consistently strengthened these correlations because they included host size and vigor. The relationships obtained can be used to assess growth responses of plantation red pine to Saratoga spittlebug attack. Shoot growth correlations can be used to evaluate growth loss directly, and the photosynthetic potential and branch vigor correlations can be used to measure lost growth potential. Reduced photosynthetic potential means less carbohydrate production and, therefore, reduced fiber production. Reducing photosynthetic potential for one year affects growth potential for at least three years-—until these needles are shed. A reduction of branch vigor as indicated by reduced bud lengths means fewer needle primordia and less potential for shoot growth the following year. Thus, the following year's photosynthetic potential is directly reduced in addition to the effects of less photosynthetic area from the current year. So, tree growth is actually influenced for at least four years. 47 1.00 r = .69I n = 78 Y = -.442 + .l36X .80- 2 .604 .401 .204 .00- PROPORTIONATE REDUCTION OF BUD LENGTH -.20 1 T . , T . . , . 1 . 2.00 3.00 4.00 5.00 6.00 7.00 8.00 LOG 0F INSECT-DAYS PER BRANCH Figure 9.--Log of insect-days per branch regressed on proportionate reduction of bud length for plot group combination P61, P62, P65, P68 and P69. 48 1.00 r = .7153 j: n = 713 I- Y = .l78 +.|52X o .80- 2 Z 8 1.1.! .J O D m .50- UL O 2 9 F- 0 .4’0‘ 2) O LIJ 0: 1‘3 .204 <12 2 9 [.— 2% a 000' O (r O. -120 v 1 1 I 1 r 1 l ' 7 V -3.00 -2.00 -l.00 0 1.00 2.00 3.00 LOG OF INSECT-DAYS PER BRANCH-UNIT Figure lO.--Log of insect-days per branch-unit regressed on proportionate reduction of bud length for the plot group combination P61, P62, P65, P68 and P69. 49 Growth loss and lost growth potential can be calculated by determining the insect exposure (i.e. insect-days) per branch resulting from a population of spittlebugs. The next sections explain how insect exposures were derived, and how these studies related to eval- uations of whole tree growth response. PREDICTION OF INSECT EXPOSURES PER BRANCH The objective of this study was to examine the relationship between insect exposure per upper-whorl branch of the growth response studies and exposures of trees to spittlebug populations. A formula developed by Ewan (1961) relates the population of late instar spittlebug nymphs to the number of feeding scars per ten linear centimeters of two-year-old branch internode. By relating this formula to the insect exposures per branch of the growth response studies, a nymphal survey can be derived which links populations of spittlebugs to tree growth responses. According to Ewan (1961) if one is given an estimate of the population of nymphs (A) and the tree-units (B) in an infested population, the number of feeding scars per ten linear centimeters of two-year-old branch internode (X) can be calculated. That is, X = K(A/B) where K = 31.3 (constant). Methods and Results To relate Ewan's formula to the insect exposures of the growth response studies (insect-days per upper-whorl branch), insect-days per branch was calculated for all third to sixth whorl branches (n = 91) of all six trees of the feeding distribution study. This was done by dividing total scars per branch by the seasonal average of 2.63 feeding SO 51 scars per day (Ewan, 1961). When the mean insect-days per branch per tree was regressed on the mean number of scars per ten centimeters of two-year-old internode per tree, a correlation of .96 (p <0.0l)(S.E. = 8.86) resulted. In order to corroborate these findings, four plots were estab- lished in red pine plantations in the summer of 1977 in areas with low to moderate spittlebug infestations. Unfortunately, severe infestations could not be found due to a widespread decrease of spittlebug populations. Four branches from the upper whorls of two trees in each plot were collected in late August after 90-95 percent of the feeding injury had occurred. These branches were peeled and scar counts made. By regres- sing the mean insect-days per branch per tree (calculated as before) on the mean feeding scar density on the two-year-old internodes per tree, a correlation of .88 (p1 xcuco oze 5.33m .msnmpu0_am emopeemm me“ Go Novas Neuwmopoom11.e_ mesmwu mun: Ik’OCO ughau >h.-_(h¢03 much magnets; to». .0604 1.0) 10.1 wadin- um w) um 1.4.x 8.— 0252...‘ 102‘: ”401’ mac; 31.0) whims! g3 uh‘twsl _ _ _ 52.33... .3 .2555 3028.3 2.33... 30.. 9.39, uptxuooauioq wage 30-. _ mo_2o .- IIIIIIIIII _ mumblhowmz. _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ 4358.. C828 _ _ 58¢. 92 _ _ 833.. 3.... mm: _ _ _ _ _ _ _ _ 3.» on. 343»: 28:32 _ _ 888.. 888.. 33: . _ £01.92. age-MWMHu 20.5550... I 292.53.. gnaw”? 283.. 5:9 N 52 its»: 9528 32.5 11 8.63.. 7 955.2. 59.93:. _ . 29:55.80 1 5.05.3 3.30: 9.28: 6» 0.536 845 93 @301!“ 67 injury. A greater density of nymphal hosts not only means a greater population potential, but also greater competition for soil moisture. This, in turn, reduces tree growthe-particularly on the moisture poor spodosols of northern lower Michigan. This means there will be smaller trees which are more susceptible to injury, and a longer time to crown closure which shades out the nymphalhosts. So, both the insect and the ground cover reduce tree growth and, thereby, increase the probability of tree injury. The population of late-instar nymphs relates directly to the population of adults (Figure 14). Further, the adult population can be estimated with a high probability from the nymphal population because there is little late-instar nymphal mortality. The adult population determines the feeding exposure to the tree and this in turn determines the density of feeding scars--relative to tree size and density. The scars make resin filled pockets which block water transport, and cause a moisture stress within and distal to the damaged areas. Scars occur all over the tree but predominate on the upper whorls where adult feeding concentrates. Stress reduces shoot and needle growth as well as future growth potential. Evidence of the latter appears as shorter needles and smaller buds. Shorter needles reduce photosynthetic potential, and smaller bud size means less needle primordia for needles the next year. These together mean less capability to produce carbo- hydrates and, therefore, wood fiber due to decrease in the amount of needle area for at least three years--the period the tree retains its needles. The immediate reduced growth means less wood volume in the 68 current year, and if reduced sufficiently it can cause tree deformity and mortality. As the upper-whorl growth decreases, resulting deform- ities increase. Deformity progresses from limbiness and large knots, to sweep, fork, and crook from lower-whorl branch dominance and hole necrosis to tree mortality. Also, as crook and sweep increases, the susceptibility to snow damage also increases. That is, trees bend excessively under snow load from the sweep and crook. An additional effect of the feeding occurs when the bole is partially girdled by necrosis from a high density of feeding punctures. Compensatory growth occurs on the opposite sides of the stem producing a sweep in the bole and structurally weakened wood. Trees often break at this point, and lower-whorl take-over results in multiple stems, crook, and/or more sweep. Several factors interact to bring about a spittlebug outbreak that results in severe host damage. Some of these factors can be man- ipulated so that spittlebug populations will be arrested before intoler- able damage occurs. These can be examined through an insect-tree pop- ulation dynamics model (Figure 15) which interlinks with the ecological model (Figure 14). It is evident that if the ground cover could be regulated, spittlebug control would follow. However, to date this approach has met with little success. Sweet-fern, which is the most important nymphal host, propagates from a rhizome-like stem. Defoliation studies indicate that due to subsequent sprouting, the crop actually increases in density the year after control. Herbicidal applications are usually unsuccessful 69 ALTERNATE HOSTS MOISTURE ”0'37”“ TEMPERATURE I EGG NYMPH. L‘\1 Low ,4 SPITTLEBUG TEMP. DYNAMICS T - PARASITES v EGG N2 - N5 / PREDATORS W‘T'm lNi (our PARASITES ADULT ______________________ 1 ' 1 ' SMALL LARGE t: : SAPLING SAPLING . - _ - _ ______________ _ _ _ _' INSECTS DISEASES = TREE ETc. DYNAMICS ‘ P0LE + SEEOLING ~——©———4 MATURE Figure 15.--Population dynamics model of the Saratoga spittlebug--red pine ecosystem. 70 because adequate dosage to kill the sweet-fern is expensive and harmful to the tree. A biological agent (defoliator, etc.) has not been found to control sweet-fern. The soil influences both tree vigor and species of ground cover and their abundance. Poor soils produce poor tree growth and generally sparse ground cover. Although planting on poorer sites might reduce ground cover abundance and therefore spittlebug risk, tree growth would probably be reduced more. More logically, better sites allow trees to attain crown closure at an earlier age and/or allow trees to surpass the spittlebug susceptable height (15 feet) sooner. Even though better Sites potentially foster higher risk ground cover associations, fewer controls would be necessary to get plantations past the spittlebug susceptible stage. Close spacing insures early stand closure. However, the current tendency is to plant 8' X 10' or 10' X 10' spacing, so closure time will be lengthened in future plantations. Besides the alternate host, there are several other factors which limit the spittlebug in its various life stages. Parasites, predators, migration, and weather are but a few. Some of these can be manipulated, and will be, after extensive population dynamics studies are made and key factors useful for management are isolated. Predictive Model Quantification of certain parameters derived from this study permits construction of a predictive model of spittlebug damage (Figure 16). Potential damage can be predicted for any susceptible red pine stand and for land proposed for planting red pine (Figure 16). To do 71 02 .xueuue manmppawam mmouegem LoN paces m>Nu0_umLN11.o~ mLzmwd I was .335 > Itin— 3le 858 mg 0233-: Guam... .28 44:. to» u g8.- nu> 83.38 «3033» ’ w .518. §(.E is u? >h.4¢h§ mg: mg ghzmocr as; 02.!th Igziucuuc “—33..th thad§ ugh to! O... whiz; g ghzmhca ’01. 3 AShzuhOa l—_LI whEIwuxh , L 72 this the area of concern Should be delimited and then risk rated. Wilson et a1. (1977) have shown that the alternate host association is useful for estimating the risk of spittlebug damage. Damage is thus based on the sum of the percentage of ground cover occupied by sweet- fern and other alternate hosts. Low risk means low damage potential and the stand is safe from the insect or fallow land is safe to plant to red pine. Should the area be rated moderate or high risk, then potentially there will be moderate to high potential damage. Unplanted land, then, should not be planted to red pine unless alternate host and/or insect control is considered in the management plans. Estab- lished plantations Should be examined further in order to predict the degree of damage. This can be done by surveying the late-instar nymphal population and determining the tree-units. Together these give nymphs per tree-unit which then can be related to the mean number of adult insect-days per upper-whorl branch by using the equation Y = -8.140 + 53.897X where Y = mean insect-days per upper-whorl branch and X = nymphs per tree-unit. Insect-days can further be related to growth loss by the equation: Proportionate reduction of growth = -.558 + .211X where X is the log of the insect-days per upper-whorl branch for one or two consecutive years accumulated feeding exposures. Yearly surveys provide more accurate means of accumulating insect exposures; however, this degree of accuracy may be unnecessary. As an alternative, the current insect exposure can be used alone as a conservative estimate of potential growth loss. Spittlebug populations in areas of moderate to heavy damage can build up rapidly, making the previous insect exposure 73 a relatively small proportion of the total exposure. 0n the other hand, the added influence of the previous feeding could be used as a subjective modifier if near-threshold values are derived from the current insect exposure. In both instances it is assumed that the population is being monitored as the population increases--before a critical damage threshold has been exceeded. The degree of growth loss directly predicts tree damage. A value of less than 25 percent growth loss Shows only reduced growth of the tree and a small loss of potential growth for the following two- three years. Percentages from 25 to 40, however, Show trees having scattered partial branch flagging and some deformity from light sweep and large lower-bole limbs. Percentages from 40 to 70 provide some trees with top kill, whole branch flagging and serious degrade from sweep, crook, multiple stems and large lower limbs. Growth loss of 70 percent or greater leads to numerous trees top killed, heavy degrade, and a few to many dead trees. These predictive values and actual results have certain socio- economic considerations which must be considered by the land manager before he can ascertain insect impact. Socio-Economic Considerations Forest lands are managed for a multiple of uses, though only one may dominate at a particular time or location. Multiple use concepts then should be considered for established and planned red pine planta- tions. The Saratoga spittlebug can modify the tree and the environment in a stand, and, therefore, can change or modify management goals. 74 The socio-economic reaction from such changes determines the impact from the insect in terms of timber, wildlife, recreation and watershed values. The quantity and quality of timber produced is a category of impact of considerable interest to the forest manager. Plantations are managed to produce an end product such as lumber, utility poles, pilings, pulp, posts, or some combination thereof. The spittlebug reduces the growth of trees, and deforms, degrades and kills them--all detriments to timber management goals. Saratoga Spittlebug damage often reduces sawlog and utility pole quality by $100-$200 per acre, and may completely degrade a stand of trees for these uses:I Once damage thresholds are surpassed the resulting degrade may render the trees unfit for many other forest products, depending upon the Size of the tree and the severity of the injury. Small trees may recover if the spittlebug population is kept under control, and the residual scars and deformities could be confined to the lower half log of the bole. However, severe injury which causes lower-whorl branch dominance and codominance leading to multiple stems, sweep, and crook could render the trees useless for utility poles, pilings, and portions of the lower log for lumber. Tree mortality, especially of small trees, could also result. Corrective pruning could help recovery of tree form, but this is costly and the benefits may not out-weigh the costs. Also, the interior scars (necrosis and stain) would still degrade the bole for lumber. Pulpwood could be considerably degraded as well. There 1Personal communication from William Stump, State and Private Forestry, U.S. Forest Service, St. Paul, Minnesota. 75 are extra loading and hauling costs in transporting crooked pulp sticks. Because whole tree chippers are not widely used, crooked pulp sticks could be culled. If the end product of many years' growth cannot be economically utilized, then there is a waste of time, energy and space. Also, growth loss from Spittlebug attack results in increased rotations and decreased wood volume. In general, the spittlebug causes a sizeable negative impact on the forest from the timber production standpoint. The watershed component of spittlebug impact seems minimal in the Lake States and probably should be considered of neutral value. Generally, the spodosols underlying a majority of these red pine planta- tions allow little run-off due to rapid penetration and infiltration of water. Also, the nymphal stage of the spittlebug requires surrounding ground vegetation for survival. Consequently, the spittlebug's damage potential is dependent upon a sufficient density of suitable ground vegetation; thus, abundant ground cover relative to initial site con- ditions remains after tree mortality to protect the soil surface from rain-drop compaction, wind erosion and aids soil stabilization. Areas with moderate to heavy damage or damage potential could be left as natural habitat for the enhancement of wildlife values or recreational-aesthetic values. The influence upon wildlife values is uncertain as yet, but an overall positive impact may result from such areas. Spittlebug mortality occurs as a pocket infestation. These pockets correspond to areas of suitable ground vegetation for nymphal survival and population build-up. The pockets are generally well vegetated and tree mortality creates "edge" within the plantation--both of which are amenable to wildlife. 76 Recreational-aesthetic values may be enhanced by providing increased variety of scenery, but are probably little affected depending upon public notice. Leaving large pockets unplanted provides less substrate for spittlebug build-up because nymphs establish primarily within ten feet of the tree (Ewan, 1961). Therefore, the decision to leave large areas natural may have a positive impact upon timber values too. Of course, increased branchiness of perimeter trees and loss of pocket area for wood production would have to be weighed against possible savings of control costs and impact upon all multiple use values. LITERATURE CITED 77 LITERATURE CITED Anderson, R.F. 1947. Saratoga Spittlebug injury in pine. J. Econ. Entomol. 40(1):26-33. Averill, R.D. 1977. Impact of redheaded pine sawfly, Neogjprion lecontei (Fitch), on young red pine plantations. _Ph.D. disser- tation, Michigan State University. 127 pp. Benjamin, D.M., H.O. Batzer and H.G. Ewan. 1953. The lateral-terminal elongation growth ratio of red pine as an index of Saratoga spittlebug injury. J. of For. 51:822-823. Campbell, R.A. 1962. A guide to grading features in southern pine logs and trees. U.S.D.A., For. Service, Southeast For. Exp. Sta., Sta. Paper No. 156. 23 pp. Clausen, J.S. and T.T. Kozlowski. 1967. Food sources for growth of Pinus resinosa Shoots. Advanc. Front. Pl. Sci., New Delhi 18:23- . Dickmann, D.E. and T.T. Kozlowski. 1968. Mobilization of Pinus resinosa cones and shoots of 014-photosynthate from needles of different ages. Amer. J. Bot. 55(8):900-906. Duff, B.H. and Norah H. Nolan. 1953. Growth and morphogenesis in the Canadian forest species. I. The controls of cambial and apical activity in Pinus resinosa Ait., Can. J. Bot. 31:471-513. Ewan, H.6. 1961. The Saratoga spittlebug: A destructive pest in red pine plantations. U.S.D.A. Tech. Bull. 1250. 52 pp. Forward, Dorothy F. and Norah J. Nolan. 1962. Growth and morphogenesis in the Canadian forest species. VI. The Significance of specific increment of cambial area in Pinus resinosa Ait., Can. J. Bot. 40:95-111. Gordon, J.C. and P.R. Larson. 1970. Redistribution of C144labeled reserve food in young red pines during shoot elongation. For. Sci. l6(l):l4-20. Guilkey, P.C. 1958. Managing red pine for poles in lower Michigan. U.S.D.A., For. Service, Lake States Exp. Sta., Sta. Paper No. 57. 21 pp. 78 79 Hanover, J.H. 1963. Geographic variation in ponderosa pine leader growth. For. Sci. 9:86-95. Jackson, W.L. 1962. Guide to grading defects in ponderosa and sugar pine logs. U.S.D.A., For. Service, Southwest For. and Range Exp. Sta., Berkeley, Calif. 34pp. Kozlowski, T.T. 1964. Shoot growth in woody plants. Bot. Rev. 30: 335-379. Kozlowski, T.T., J.H. Torrie and P.E. Marshall. 1972. Predictability of shoot length from bud size in Pinus resinosa Ait. Can J. For. Res. 3(1):34-38. Kozlowski, T.T. and R.C. Ward. 1961. Shoot elongation characteristics of forest trees. For. Sci. 7(4):357-368. Kozlowski, T.T. and C.H. Winget. 1964. The role of reserves in leaves, branches, stems and roots on shoot growth of red pine. Amer. J. Bot. 51(5):522-529. Kulman, H.M. 1965. Effects of artificial defoliation of pine on sub- sequent shoot and needle growth. For. Sci. 11:90-98. Larson, P.R. 1962. The indirect effect of photoperiod on tracheid diameter in Pinus resinosa. Amer. J. Bot. 49(2):132-l37. 1963. The indirect effect of drought on tracheid diameter in red pine. For. Sci. 9(1):52-62. Mamaev, S.A. 1956. The moisture content of Shoots and conifer needles and its effect on branch growth. (Translated from Russian) IPST Cat. No. 5216. U.S. Department of Commerce, Clearinghouse for Fe. Aci. and Tech. Infor., Springfield, Va. 3pp. Marion, B.H., J.V. Berglund and A.L. Leaf. 1963. Morphological and chemical analysis of red pine (Pinus resinosa Ait.) buds. Pl. Soil 28:313-324. Miller, W.E. 1965. Number of branchlets on red pine in young planta- tions. For. Sci. ll(1):42-49. Pharis, R.P. 1976. Probable roles of plant hormones in regulating shoot elongation, diameter growth and crown form of coniferous trees in TREE PHYSIOLOGY AND YIELD IMPROVEMENT edit. by Cannell, M.G.R. and Last F.T. Academic Press. pp. 291-306. Sucoff, E. 1971. Timing and rate of bud formation in Pinus resinosa. Can. J. Bot. 49(10):1821-1832. 80 Szymanski, S. and Szczerbinski, W. 1962. (Trans. from Polish). Buds as indicators of shoot vigor. Rocznik Sekeji Dendrologiczej Polskiego Towarzystwa Botanicznego V0. 10:275-304. USDA. 1972. Insect and disease impacts on forest resource uses. values, and productivity needs and opportunities for an integrated research and development program. USDA, For. Ser. 76 pp. (mimeo). Wilson, L.F., G.C. Heaton and P.C. Kennedy. 1977. Development and survival of Saratoga spittlebug nymphs on alternate host plants. The Great Lakes Entomol. lO(3):95-lO5. T ARIES MICHIGAN s RTE UNIV. LIBR 1111111111“ I1"11111111111111 3129 066490 $3011 65