p >‘5’-3). t‘.'.' e‘. 5 1."'-‘-" I'tr . -o ._ IIIIvIIflIquIwIdt-‘z naueziriafiwzr GENETIC VARIATION OF SEED AND SEEDLING CHARACTERISTICS OF PAPER BIRCH (BETULA PAPYRIFERA) AND EUROPEAN WHITE BIRCH (BETULA VERRUCOSA) Thesis for the Degree of M. S. MICHIGAN STATE UNIVERSITY MICHAEL FRANCIS SOHASKY 1977 ABSTRACT GENETIC VARIATION OF SEED AND SEEDLINGS CHARACTERISTICS OF PAPER (BETULA PAPYRIFERA) AND EUROPEAN WHITE (BETULA VERRUCOSA) BIRCH By Michael Francis Sohasky The key to the improvement of paper birch (Betula pabygifera) is a knowledge of how much genetic variation present in the species. Because of the growing importance of paper birch as a timber producing tree and its potential for artificial regeneration, information on geographic var- iation patterns in growth rate and other desirable traits is an important first step towards its improvement. The objectives of this study was to investigate the amount and pattern of genetic variation in paper birch through the measurement of l) germination capacity and germinative energies of the seed, 2) growth rates, and 3) develOpmental and morphological characters of juvenile seedlings. Variation patterns in the germination characteristics of paper birch seed were investigated on 219 half—sib families collected in Michigan, and 25 seedlots of Finnish birch. Studies of growth rates, and developmental and Michael Francis Sohasky morphological characters were done on container-grown seedlings grown under greenhouse conditions. Temperature and geographic origin each had a significant effect on the germination capacity and germinative energy of paper birch seed. Seed from the Upper Peninsula (UP) of Michigan revealed a wider range of temperatures Optimum for germination, and increased germinative energy, in comparison with seed from the Lower Peninsula (LP). There were distinct differences in the initiation phase of seed germination, with UP families beginning first. There was also genetic variation between- and within- populations of paper birch in germination capacity and germ- inative energy. Seed from populations in the western UP began germination sooner, proceeded more rapidly, and reached optimum levels earlier than pOpulations in the eastern UP and LP. Genetic variation in juvenile height growth of container-grown paper birch seedlings grown under green- house conditions was essentially random with respect to major geographical gradients. Most of the variation in height growth of 4- and 5-month-old greenhouse grown seed- lings was due to within-population differences (87%). Thus variation between half-sib families was greater than dif- ferences among widely separated pOpulations. The amount of variation in growth between families differed among the populations sampled but the maximum variation between any Michael Francis Sohasky two populations was small (20%) compared to the maximum variation observed between two families (68%). Comparative analysis between paper birch and European white birch (Betula verrucosa) from Finland, grown under greenhouse conditions, revealed the height superiority of the Finnish sources of birch. If this superiority holds true in test plantations it would indicate a possible use of European white birch for reforestation purposes in Michigan. Variation in 26 develOpmental and morphological characters measured was essentially random over the natural range of paper birch, following no geographical or clinal pattern. Significant genetic variation was found in a number of characters, including number of lateral branches, stem pubescence, leaf number, leaf area, leaf length, and leaf shape. Variability in these characters exists both between.widely separated populations of paper birch, and between paper birch and European white birch. GENETIC VARIATION OF SEED AND SEEDLING CHARACTERISTICS OF PAPER BIRCH (BETULA PAPYRIFERA) AND EUROPEAN WHITE BIRCH (BETULA VERRUCOSA) BY Michael Francis Sohasky A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Forestry 1977 ACKNOWLEDGMENTS I wish to extend my sincere gratitude to my major professor Dr. James W. Hanover, whose guidance and generous assistance were essential to the successful completion of this study. My sincere appreciation goes to the other members of my guidance committee--Drs. D.I. Dickmann and S.N. Stephenson-~for their valuable contributions. I also wish to thank all my colleagues for their kind assistance in various aspects of this study, especially Mr. Bruce Bongarten for his valuable assistance and statistical counsel. Finally, I would like to thank my wife, Pam, for her sacrifices and encouragement during the preparation - of the thesis, and my entire academic career. ii TABLE OF CONTENTS LIST OF TABLES . . LIST OF FIGURES . . . . CHAPTER I. Introduction . . . . Species Description . . . Objectives . . . II. Geographic Variation in Germination of Birch Seeds . . . . . Introduction . . . Materials and Methods . . Results . . . . . Discussion . . . Summary and Conclusions III. Geographic Variation in the Growth Rates of Seedlings . . . . Introduction . . Materials and Methods . Results and Discussions . Summary and Conclusions . IV. DeveIOpmental and Morphological Characters of Seedlings . . . . Introduction . . Materials and Methods . Results and Discussion . Summary and Conclusions . SUMMARY, CONCLUSIONS AND RECOMMENDATIONS . BIBLIOGRAPHY . . . . . iii Page iv vi and»: LIST OF TABLES TABLE Page 2—1. Analysis of variance between ten seedlots of paper birch in response to four temperature treatments. . . . . . . . . . . . . 23 2-2. Influence of temperature and seedlot on the germination of Betula papygifera seeds after 21 days 0 O I O I O O O O O O I 0' O 24 2-3. Number of days required for Betula a rifera seeds to reach 75 percent germination Igerm— inative energy) at various temperatures. . . 26 2-4. Analysis of variance between seven seedlots of paper birch in germination response to four temperature treatments. . . . . . . . . 30 2-5. Influence of temperature and seedlot on the germination percentage of Betula papyrifera seeds after 21 days. . . . . . . . . . 31 2-6. Number of days required for Betula papyrifera seeds to reach 75 percent germination at various temperatures. . . . . . . . . . 33 2—7. Average percent germination, and the range of germination means, of 25 pOpulations of Betula papygifera seeds from Michigan, and three groups of Finnish sources of birch, after 42 days. . 40 2-8. Cumulative percent germination over time for Betula papyrifera and three groups of Finnish birCh seeds 0 O I I O I O C I O I O I 42 3-1. Four-month relative heights of 25 paper birch seed sources eXpressed as percent of the overall mean (EXperiment I). . . . . . . . . . 73 3-2. Four-month height of the six best and six poor- est families of all 66 families represented in all five replicates. . . . . . . . . . 75 iv TABLE Page 3-3. Five-month relative height growth and origin data of 42 paper birch populations, expressed as percent of the overall mean. . . . . . 78 3-4. Analysis of variance of juvenile height growth of 42 paper birch seedlots 5—months from seed. 79 3-5. Correlation between juvenile height growth of 42 paper birch seedlots and environmental variables of the seedlots. . . . . . . . 79 3-6. Species, origin, and 4-month relative height growth expressed as percent of the overall mean for 8 populations of paper birch collected in 1975, 3 pOpulations re-collected in 1976, three range-wide cpulations, and two other species of birch IExperiment III). . . . . 84 3-7. Analysis of variance of juvenile height growth of 48 paper birch families 4-months from seed. 85 3-8. Anal sis of variance of juvenile height growth of 7 seedlots of birch -months from seed. . 87 3-9. Correlation analysis between juvenile height growth of paper birch progeny and environmental variables at the seed origin. .. . . . . . 89 4-1. Developmental and morphological characters measured on 16 seedlots of paper and European white birch. Each value is a mean of 20 seedlings. . . . . . . . . . . . . 104 LIST OF FIGURES FIGURE Page l-lI The range Of paper birCh. I I I I I I I 2 2-1. Location and population numbers of the 25 paper birch populations. . . . . . . . 13 2-2. Location and identification numbers of the ten seedlots (families) used in Experiment I. 16 2-3. Location and identification numbers of the seven seedlots (families) used in Experiment II. 20 2-4. Cumulative percent germination of ten combined seedlots of paper birch from the Up er Pen— insula (UP) and Lower Peninsula (LP of Michigan, at temperatures of 21° and 320C. . 28 2-5. Cumulative percent germination of seven combined seedlots of paper birch from the UPOand LP ofoMichigan<5 at temperatures of 27 , 29 , 32 , and 35 C. . . . . . . . 35 2-6. Cumulative percent germination of all seven individual seedlots of paper birch, avgraged ov8r all three temperatures of 27 , 29 , and 32 CI I I I I I I I I I I I I I I 37 2-7. Cumulative percent germination of UP and LP combined populations over a 42 day period. . 44 3—1. Location and population numbers of the 25 paper birch populations used in Experiment I. 59 3-2. Location and population numbers of the 42 paper birch populations used in Experiment II. . . 64 3-3. Location and pOpulation numbers of the 8 paper birch populations from Michigan used in Exp- eriment III. Populations marked with a asterisk are represented by seed collected in 1975 and 1976. I I I I I I I I I I I I I I 68 vi vii FIGURE Page 4-1. Location and pOpulation number of the 12 paper birch populations used for this study. . . . 97 4—2. Diagrammatic sketch of a leaf showing the linear measurement numbers. . . . . . . . lOl CHAPTER I INTRODUCTION Species Description Paper birch (Betula papypifera Marsh.), also known as white or canoe birch, is one of the 40 species of trees and shrubs that comprise the genus Betula. Paper birch has a transcontinental range, extending northward almost to the absolute limit of tree growth in the Boreal Forest Region of Canada and Alaska (Figure 1-1). It extends continuously across Canada from Newfoundland to Alaska, and penetrates the United States in the Northeast, southward to Pennsylvania, in the Lake States, and in northern Idaho and Montana. Out- liers occur as far south as North Carolina, West Virginia, Iowa, Nebraska, in the Dakotas, and Colorado. Paper birch reaches its best deve10pment in the northern New England area (Hutnik and Cunningham, 1961; Fowells, 1965). In eastern North America, paper birch shares the cool- er part of its range with B. cordifolia, and in the warmer parts its range overlaps with B. populifolia. In the northern prairies and British Columbia paper birch is sympatric with B. neoalaskana and overlaps in Alberta and British Columbia with B. fontinalis, with which it some- times hybridizes (Dugle, 1966; Grant and Thompson, 1975). Figure 1-1. The range of paper birch. .583 .89!— HI can: 2—H. L, In paper birch, Brittain and Grant (1966, 1967) have found somatic chromosome numbers of 56, 70, and 84, possibly attributed to hybridization. Paper birch is a short-lived, fast growing tree. Trees reach maturity as early as 60 years of age, with a maximum age of about 140 years. It attains heights of 50 to 70 feet and average stem diameters of 12-24 inches. The bark at first is dark brown or bronze, but soon turns chalky to creamy white, separating into thin, papery strips thereby giving it its distinctive name. Leaves are oval to ovate, commonly 2 to 3 inches long, and 1% to 2 inches wide. Leaf margins are singly to doubly serrate with the base being rounded or obtuse. Staminate aments are i to 1% inches long, usually borne in 2's or 3's. The cone-like pistillate catkins are 1 to 1% inches long, slender-stalked and pendant each scale may bear a single small, winged nut (seed), which is oval in shape with two persistent stigmas at the apex (Harlow and Harrar, 1969; U.S. Forest Service, 1974). Paper birch is a cold-climate species, growing in areas characterized by short cool summers and long cold winters with long periods of snow on the ground. Paper birch grows under an extremely wide range of habitat conditions. It is found as far north as the 55°F July isotherm, with out- pockets extending almost to the absolute limit of tree growth (Halliday and Brown, 1943). It seldom occurs nat— urally where the average July isotherm exceeds 70°F (Fowells, 1965). It occurs in areas where killing frosts can occur in 5 the spring as late as the first week in June, and in the fall as early as the first week in September (Marquis et al., 1969). In view of its transcontinental range, paper birch can tolerate wide variations in the patterns and amounts of precipitation. The average annual precipitation within the commercial range varies from 25 to 45 inches, with about half occurring as snow. Paper birch is confined largely to soils of the Spodosol region. However, it occurs on gray-brown and brown podzolic soils. Within these groups, numerous soil series and soil types are satisfactory for the growth of paper birch. Maximum development is reached on well-drained sandy loams. It is common on shallow stony soils and even occurs on wet bog and peat soils. The soils range from acid to highly calcareous, with the optimum pH in the range of 4.5-6.0. For the most part, paper birch occurs at all elevations, and on all aspects and lepes. It is one of the few hard- woods found near timberline on the highest mountains in New England and New York. However, in the southern part of its range, it occurs only on the cooler sites-~at higher elevations and on steep north- and east-facing s10pes (Fowells, 1965). Paper birch is an intolerant tree, occurring as a scattered tree in the mixed coniferous-hardwood forests of the north. They are less tolerant than their common forest associates—-principally sugar maple, beech, yellow birch, red spruce, white spruce, and balsam fir. For this 6 reason paper birch generally disappears from a stand after one generation, leaving only the more tolerant species. In some localities, especially the northern Lake States, paper birch together with white spruce and balsam fir, comprises a large portion of the forest (Hyvarinen, 1968). Fire plays an important role in the regeneration of paper birch. After fire, paper birch often seeds in large areas where mineral soil has been exposed, especially on moist sites where it may form nearly pure stands. In this respect, paper birch is similar to the aspens and is com- monly found in mixture with them on old burns (Harlow and Harrar, 1969). Reproduction is mostly from seed, although moderate stump sprouting can occur, especially after cutting or fire. Paper birch growing stock volume in North America is estimated to be about 36 billion cubic feet. Of this, about one-fifth is in the U.S. and four—fifths in Canada. Twenty-seven percent of the 7 billion cubic feet of grow- ing stock in the U.S. is located in the Lake States, 43 percent in Alaska, 29 percent in the Northeastern States, and 1 percent in the Pacific Northwest (Quigley and Babcock, 1969). The approximately 1.6 billion cubic feet represents about one-fourth the amount of the aspen volume in the Lake States (Hyvarinen, 1968). Pulpwood is the most substantial use for paper birch in the Lake States. In 1965 pulpwood production was around 55,000 cords (Horn, 1966). In 1974 however, 184,000 7 cords of paper birch pulpwood was cut in this region (Blyth and Hahn, 1974)- Fuelwood is the second most important use of paper birch, especially for small—sized sticks. The largest paper birch trees are out into sawlogs and veneer logs, with the highest quality logs used for veneer. The wood of paper birch is light, strong, smooth textured, and straight grained. A considerable quantity of paper birch is used for miscellaneous turning products, especially in New England (Hyvarinen, 1968). Paper birch is also a valuable tree aesthetically. Its showy bark makes paper birCh an important tourist attraction in the northern landscape, where recreation is rapidly growing. Paper birch has a number of important disease and insect pests that cause widespread economic losses. The agents most damaging economically to paper birch are micro- organisms that cause discoloration and decay. The principle pathogens are Torula ligpiperda causing red heart, Eppgg igpiarius and Pppig obligua associated with heart rot (Hepting, 1971) . A Major insect pests associated with paper birch include the bronze birch borer, Agpilus anxius, birch leaf miner, Fenusa pusilla, and the forest tent caterpillar, Mglgppgppg disstria (Conklin, 1969). In the 1940's and early 50's widespread and serious decline of birch, known as "birch dieback", occurred in the 8 Northeast and Canada. Paper birch is also susceptable to a condition known as post-logging decadence, where paper birches have been exposed by opening the stands (Fowells, 1965). Objectives Because of its rapid growth, high value wood, and aesthetic value, the demand for paper birch will continue to increase, especially high—quality veneer and sawlogs. To help meet this increase in demand, the establishment of plantations containing genetically improved progeny superior in growth rate, quality, and form is essential. To effectively meet this demand by tree improvement, the extent of genetic variation must be estimated. While this information is known for many important commercial trees such as eastern white pine, jack pine, red pine, and white spruce, research on paper birch is scanty. The objective of this study was to investigate the amount and pattern of genetic variation of paper birch (Betula papypifera). In addition, European white birch (Betula verrucosa) was included for comparative analysis between the two species in: l. Germination capacity 2. Germinative energy (rate) 3. Growth rates of seedlings 4. Developmental and morphological characters of seedlings CHAPTER II GEOGRAPHIC VARIATION IN GERMINATION OF BIRCH SEEDS Paper birch (Betula papyrifera) seed ripens from early August till mid-September. Seed dispersal begins soon after ripening and continues throughout the winter months. The seeds are wind-disseminated, and because of their light weight they can be carried long distances. However, most seeds fall within a few hundred feet of the parent tree (Marquis et al., 1969). Due to their very small size, birch seeds are very sensitive to the light, moisture, and temp— erature conditions at the time of germination. The effect of light and temperature on paper birch seed germination has been demonstrated by many workers. Vaartaja (1952) and McDermott (1953) found that the germ- inating capacity of Betula pubescens and B. pigpg seeds is clearly stronger in light than in darkness. Studies by Black and Wareing (1954, 1955) show that the germination capacity of unchilled B. pubescens seeds depends on both the light period and temperature. Under long daylengths and a temperature of 15°C, germination is high (90%), whereas it is low under short-day conditions (25%). At 20°C, the photoperiodic response disappears and the percent- age of germination is high under both long- and short-day 10 conditions. Cold treated B. pubescens seeds do not have any light requirement for germination (Black and Wareing, 1954). Vaartaja (1959) has shown photoperiodism with respect to the germination of B. pubescens seeds, and also suggested that the response of birch seed germination to light does not vary greatly among seedlots of different origins. Probably the most crucial factor in determining germ- ination of birch seeds in the soil is a suitable temperature. As seeds are the principle means of propagating the species, their germination should occur at a time when it will favor the survival of the young seedlings. In many species, including paper birch, this time does not coincide with its dispersal. Despite the fact that moisture is often present when the seeds are shed, the seeds may not germinate. Paper Ibirch seeds may have a minimum.temperature requirement which delays germination. The value of such a temperature requirement is obvious. Due to the large variations in climatic conditions found within the natural range of paper birch, it is possible that geographic variation patterns in germination capacity and germinative energy have evolved in relation to the temperature conditions at their origins. Reports of variation in germination of seeds from different geographic origins in response to temperature has been contradictory for a number of species. McNaughton (1966) reported that Typha populations from southern origins germinated at lower temperatures. Fraser (1971) observed 11 just the Opposite effect with white spruce provenances, i.e., northern provenances of white spruce germinated at lower temperatures than those from warmer more southerly prov- enances. Mergen (1963) reported that seed of eastern white pine from northern sources germinated more rapidly than southern ones. More recently, Winstead (1971) reported that northern U.S. pOpulations of sweetgum require the longest time to reach peak germination at all temperatures, whereas Mexico seedlots showed earliest peak germination. In white spruce, Fraser (1971) again found just the reverse, with northern provenances germinating sooner and ranking higher than southern provenances at temperatures up to and including 60°F. In birch, Govorukha and Mamaev (1971) were the first to study the effect of geographic origin on the germination of B. verrucosa and B. pubescens at various temperatures. The germination of birch seeds at various temperatures, collected in the steppe, S. tiaga, and N. tiaga zones of Russia, revealed a clear relationship to geographic origin. Seeds of both birch species from the more southern (steppe zone) origin, were found to have a wider temperature range and optimum, as well as increased germinative energy in comparison with birch seed from the more northern (N. tiaga) origin. Higher temperatures were required for the germination of seeds from the northern origin. Germ- ination of seeds of northern origin also decreased sharply with an increase in temperature. Germinative energy or 12 speed of germination varied from 95 percent germination in 5 days at 30°C for the southern origin to 95 percent germ- ination in seven days for the more northern origin. At temperatures of 200 and 35°C, even wider differences in germinative energy were observed. Knowledge of geographic variation in seed germination of paper birch provenances in response to varying temp- eratures is non-existent in paper birch. The objective of this study therefore was to measure the variation in germ- ination characteristics of different seed sources of paper birch in response to temperature. Materials and Methoqg The seeds used in this study originated from a state- wide collection of fruiting catkins (strobili) of paper birch during September, 1975. Collections began in the western portion of Michigan's Upper Peninsula (UP) on September 3rd and ended on September 25th in southern Michigan (LP). The seed sources used here are part of a larger collection described later. Collection areas (populations) were uniformly dist- ributed over the State of Michigan.) Twenty-five different populations were collected (Figure 2—1). In each population, encompassing about a hectare in size, seeds were collected from 5 to 11 parent trees (seedlots), with an average seed- lot per population of 9. Thirteen populations and 124 seedlots in the UP and 12 populations and 95 seedlots in 13 Figure 2-1. Location and pOpulation numbers of the 25 paper birch pOpulations. l4 ION MI! I ' am (on JIM III Tu ran unau- ‘srcuu I m . cam ”(“73 nun/mt 'uuu'v? — ”MRI“! museumguul moot . I a " ,.. ,- v o. . . .l’. ,’ . /' mum ‘ a .' uwxn- ‘ L 15 the LP were collected. Each parent tree was identified and assigned a Michigan State University accession number. The number of catkins collected depended on the size of the seed crOp. For each tree, green catkins were put into paper sacks, numbered, and allowed to dry at room temperature for approximately ten days to a moisture content of four percent. Seeds, bracts, and a small amount of drierite were then placed in sealed zip-lock plastic sand- wich bags, labeled and placed in cold storage (4°C). Experiment I To study germination in response to different temp— eratures, ten widely scattered populations were chosen, five from the UP and five from the LP. A seedlot was randomly selected from each of these pOpulations. The locations and identification numbers of the selected seed- lots are shown in Figure 2—2. On October 9, 1975, the seeds were placed into covered Petri dishes containing one layer of white Whatman 9.0 cm filter paper. Distilled water was added to moisten the filter paper and was periodically applied as necessary. The dishes were placed in small illuminated, forced-draft, constant-temperature Sherer model CEL-25—7-HL growth chambers. Continuous illumination was provided by a mixture of cool-white flourescent tubes and 25 watt incadescent bulbs providing 800 ft-candles of light. 16 Figure 2-2. Location and identification numbers of the ten seedlots (families) used in Experiment I. 1? mm mum “"0 I. WI @101“: 0“!” “All ll‘fl’l “4“!“ “100' . Mr W “can .. gr ‘. .. . “:1 W. 1:5" an“: em: (Liam Mun- ”an (MOD WM #— Q mum mum umur I I (AW Jam. “HIM mm! — SIM-hair‘s“ mum In“; L 18 The eXperimental design consisted of three replications of approximately 60-75 seeds from each of the ten seedlots, in each of the four temperature treatments: 10°, 21°, 32°, and 43°C . The germinated seeds were counted daily throughout a period of 21 days. Most germination had taken place at the end of 21 days except a few seed from a southern seedlot which germinated up to 35 days. A seed was considered germ- inated when the radicle protruded 1 mm or more from the seed coat. The seeds were not stratified before treatment and were not subjected to any cold storage. Immediately after the initial drying phase the seeds were placed into the temperature treatments. Percent germination data were transformed to degrees by the arc sine (arc sine/percent germination) method to normalize the data, and they were then subjected to analysis of variance. Duncan's (1955) multiple range test was used to test significance of treatment and seedlot means in germination capacity. Seed germination capacity was calculated as the percent of seed germinating after 21 days. Germinative energy was calculated as the number of days required for 75 percent of the seeds to germinate. Bxperiment II To further study germination responses to temperature, a much narrower range of temperatures was selected. Rather 19 than using all the seedlots in Bxperipent I, three of the seedlots that revealed poor germination were excluded, leaving seven seedlots, three from the UP and four from the LP. The location and identification numbers are represented in Figure 2-3. On October 28, 1975 the birch seeds were removed from the cold room (4°C) after storage in dry sandwich bags for three weeks, and subjected to the same method of germ- ination and experimental design as used in Experiment I. Four temperature treatments used were 27°, 29°, 32°, and 35°C. The germinated seeds were counted daily throughout the 21 day period. Analysis of variance was performed on transformed percent germination data. Bxperiment III Materials for this study consisted of all 219 seedlots from the 25 populations of paper birch. In addition, three different groups of birch seed from Finland were used, through the cooperation of Martti Lepistg of the Foundation for Forest Tree Breeding, Helsinki, Finland. The three groups consisted of 1) ten seedlots from a commercial seed orchard of B. verrucosa originating from southern and central Finland, 2) five seedlots from their greenhouse intraspecific hybrids of B. verpucosa, and 3) ten inter— specific hybrids between B. verrucosa and other species of birch. 20 Figure 2-3. Location and identification numbers of the seven seedlots (families) used in Experiment pp. 21 22 This eXperiment was carried out at a constant temp- erature (30°C) for all 219 seedlots and was designed to test seedlot differences. On November 7, 1975 the birch seed was removed from cold storage (4°C) after storage in dry plastic sandwich bags for 4 weeks, and placed into covered Petri dishes in a growth chamber (model CEL 512-37) with lighting as in the previous experiments. Each germination dish contained 100-175 seeds. The germination period extended to 42 days. Germinated seeds were counted from the day germination started. This counting was done every day the first three days, and then periodically every 2, 3, and 4 days as time progressed. ‘At each observation the counted seeds were removed. Germination capacity was calculated as the percent of seed germinating in 42 days. Germinative energy was expressed as the number of days to a certain percentage of the total germination. Results Experiment I Analysis of variance of the transformed percent germ- ination data shows that temperature and seedlot each have a significant effect on the germination of Betula papypifera seeds (Table 2-1). The interaction of temperature x seed- lot suggests that the response of birch germination to 23 Table 2-1. Analysis of variance between ten seedlots Of paper birch in response to four temperature treatments. Source Of variation d.f. M.S. F-value Temperature 3 1770.6 17.9** Seedlot 9 548.0 5.6** Temperature x seedlot 27 98.8 10.4** Error 80 9.5 Total 119 ** P<0.01 temperature varies greatly among different seedlots. Germination capacity is significantly affected by temperature, as shown in Table 2-2. The optimum temperature of 32°C produced an average germination percent of 31.3. At the lowest temperature treatment of 10°C, germination capacity was significantly reduced, averaging 6.9 percent. The UP seedlots had germination in all five seedlots, where- as the LP seedlots had only two with any germination at all. Conversely, at the maximum temperature of 43°C, the LP seed- lots had no germination, but in the UP two seedlots showed some germination. At 21°C the 16 percent germination percentage is mark- edly less than the 29 percent germination at 32°C in all five seedlots from the LP. The UP seedlots showed nearly identical germination capacity at temperatures of 21°C and 32°C. Therefore, it appears that LP seedlots have a much narrower range of temperatures Optimum for germination. 24 Table 2—2. Influence of temperature and seedlot on the germination of Betula papypifera seeds after - 21 days. Seedlot Michigan Temperature 0° Seedlot NO. County 100 218* 32U 43“ Means I ------ percent ------ Upper Peninsula: 1 4- 4 Ontonagon 0.7 8.0 9.3 0.0 4.5ef -17 Keweenaw 48.1 95.1 92.4 0.7 59.1a -35 Iron 3.1 15.2 24.3 0.0 10.7cd -7O Marquette 8.8 35.7 34.7 0.0 19.8b -86 Schoolcraft 3.2 9.2 7.5 2.3 5.6de MEEEI’ 1278 '3276 3376 '576 ‘I979 Lower Peninsula: 4-125 Emmet 0.0 6.1 8.4 0.0 3.6ef -157 Presque Isle 1.2 29.3 44.6 0.0 18.8bc -196 Missaukee 0.0 0.7 3.8 0.0 ' 1.0f -209 Arenac 0.0 1.7 29.7 0.0 7.9ef -216 Sanilac 3.4 42.1 58.6 0.0 26.0b MEEE— -ITO TOTO 2975 '_575 III? Combined means: 6.9a 24.3b 31.3c 0.3d ;/' Any two means with the same letter are not sign- ificantly different (P<0.05) by Duncan's Multiple Range Test. 25 Table 2-2 also illustrates the broad variation in germination capacity of each individual seedlot. Overall germination percent of the seedlots ranged from 59.1 percent for seedlot 4-17 to 1.0 percent for seedlot 4-196. Within a seedlot, increasing temperatures up to 32°C generally increased germination, especially in the LP where it occurred in all seedlots. There was no significant decrease in germination response of northern origins (UP) at 32°C as reported for B. pubeggens by Govorukha and Manaev (1971). They reported that a temperature of 33-35°C was optimum for seed germination of B. pubescens from the southern steppe zone Of Russia, whereas the more northern origins showed a noticeable decrease in germination capacity. The influence of temperature on the rate of germination or germinative energy of the ten seedlots is evident in Table 2—3. As temperature increased rate of germination increased. At a temperature Of 32°C all the seedlots showed maximum germinative energy. Germination began soon- est proceeded at a more rapid rate, and reached Optimum levels soonest at that temperature. At the lowest temperature (10°C) not only was germ- inative energy significantly reduced, but germination was delayed. Germination began later and proceeded at a much lower rate than at the other temperatures. This experiment revealed that germinative energy of birch seeds from the northern populations (UP) was much faster than southern populations (LP) at 21°C and at the 26 Table 2-3. Number of days required for Betula a ifera seeds to reach 75 percent germinatigfig%§e;m:_ inative energy) at various temperatures. Seedlot Michigan Temperature °C NO. County 100 210 320 430 ------ NO. days - - - - - Upper Peninsula: 4- 4 Ontonagon 12.0 13.4 11.8 - i/ -17 Keweenaw 12.5 7.5 6.6 9.0 -35 Iron 14.3 8.7 5.6 - -70 Marquette 15.3 12.6 7.4 - -86 Schoolcraft 10.6 8.1 4.9 8.0 Lower'Peninsula: 4-125 Emmet - 13.8 7.6 - —157 Presque Isle 13.0 15.2 9.9 - -196 Missaukee - 17.0 - —209 Arenac - 11.4 12.9 - -216 Sanilac 10.3 15.6 12.8 - UP combined 12.2 10.1 7.5 8.6* LP combined 16. 0* 14. 3 11.4 .- l/' No germination occurred. * Represented by only two seedlots. 27 Optimum temperature of 32°C (Figure 2-4). Thus, the combined UP seedlot means show that at the Optimum temperature, 75 percent germination was Obtained in just 7.5 days, whereas the LP combined seedlot means reached 75 Percent germination in 11.4 days, a difference of 3.9 days. Seedlots from the UP showed a much faster initial rate of germination. For example, at 32°C the germinative energy of birch seeds from the UP was 51 percent in 5 days, while in the LP, it was only 6 percent. At the cooler temperature of 21°C, the LP had no seed germination until the 7th day, whereas the UP seedlots had begun germination by the 5th day. Bxperiment II This experiment was similar to E eriment I, except that a much narrower range of temperature treatments was selected to determine the optimum germination temperature for paper birch seeds, and to Obtain any further insights into variation patterns Of germination capacity and germ- inative energy. Analysis of variance of the percent germination data presented in Table 2-4 revealed that temperature and seed- lot each have a significant effect on the germination of paper birch seeds. The significant interaction of temp- erature x seedlot again suggests that the response Of birch seed germination to temperature varies among different seed- lots. 28 Figure 2-4. Cumulative percent germination of ten combined seedlots of paper birch from the Upper Pen- insula (UP) and Lower Peninsula (LP) of Michigan, at temperatures of 21° and 32°C. 100 90 80 70 60 c: .9 , 2 E50 0 (D #40 C (D 8 a? 30 G) .5 I 2 I 320 .' 3 I U I 10 I. 0 ’a“—‘ 4 5 6 29 ""'" Upper Peninsula --- Lower Peninsula 0 32°C OZHOC 8 91111213141517 Days 30 Table 2-4. Analysis of variance between seven seedlots of paper birch in germination response to four temperature treatments. Source Of variation d.f. M.S. F-value Temperature 3 1043.7 18.10** Seedlot 6 590.9 10.25** *1- Temperature x seedlot 18 57.7 3.02 Error 56 19.1 Total 83 ** P<0.0l The results from this experiment are shown in Table 2-5. The optimum temperature for germination ranges from 27° to 32°C, combined over all seven seedlot means. No significant differences existed between the three temp- erature treatments of 27°, 29°, and 32° using combined seedlot means, but the Optimum temperature for the more southerly (LP) was shifted in the direction of the lower temperature. At the maximum temperature of 35°C, the LP seedlots revealed almost no germination, with only one seedlot germinating at a much reduced capacity. Conversely, the more northerly (UP) seedlots all showed more germination than the LP seedlots at 35°C, but this temperature still gave lower germination than the other three temperatures. This is exactly the opposite the findings of Govorukha and Manaev (1971) on B. pubescens who reported that more south- erly population had 90 percent germination at 35°C and the more northerly population showed a decrease in germination to 60 percent. 31 Table 2-5. Influence of temperature and seedlot on the germination percentage of Betula papypifera seeds after 21 days. Seedlot Michigan Temperature CC Seedlot No. County 270 290 320 350 Means ------ percent - - - - - Upper Peninsula: A/ 4-17 Keweenaw 94.6 95.7 96.5 32.0 79.7a ~35 Iron 19.4 19.6 22.7 19.3 20.3de -70 Marquette 35.9 40.6 34.5 18.4 32.4c Mean 49.9 52.0 51.2 23.2 44.1 Lower Peninsula: 4-125 Emmet 14.2 8.6 6.8 0.0 7.4f -l57 Presque Isle 35.7 39.8 43.3 3.5 30.6cd -209 Arenac 39.8 22.2 17.8 0 0 20.0e -216 Sanilac 63.2 69.4 59.5 0.0 48.0b Mean 38.2 35.0 31.9 0.9 26.5 Combined means: 43.3a 42.3a 40.2a 10.5b ;/ 'Any two means with the same letter are not sign- ificantly different (P<0.05) by Duncan's Multiple Range Test. 32 As indicated in E eriment I, Table 2-5 also illustrates the broad variation in germination capacity of each individ- ual seedlot. Final mean germination capacity ranged from 79.7 to 7.4 percent, for seedlots 4-17 and 4-125, respectively. Within a seedlot, increasing the temperature at these closely spaced temperatures did not significantly increase the germ- ination. Some seedlots increased slightly with increasing temperatures, while others decreased. Germinative energy, or rate of germination, of paper birch seeds showed variation due to seedlot, temperature, and peninsula. The influence of temperature on the rate of germination of the seven seedlots is evident in Table 2-6. In contrast to E eriment I, increasing temperatures at these closely spaced treatments did not necessarily increase the germinative energy Of the seedlots. Within a seedlot or peninsula the differences are small and quite variable, with no observable patterns except for 35°C where germinative energy was significantly reduced. Between the two peninsula's, substantial differences are observed in germinative energy. Germinative energies of birch seed from the UP were much faster than the LP seedlots at all temperatures (Table 2-6). At 32°C the combined UP seedlot means revealed that 75 percent of the germination was obtained in 6.9 days, whereas the LP combined seedlot means reached the same level of germination in 13.0 days, a difference of 6.1 days. 33 Table 2-6. Number of days required for Betula papypifera seeds to reach 75 percent germination at various temperatures. Seedlot Michigan Temperature °C Combined No. County 0 o o 0 Temp. 27 29 32 35 Tmts. ------ No. days - - - - - Upper Peninsula: 4-17 Keweenaw 6.0 6.0 6.5 11.5 ‘35 Iron 6IO 9IO 7IO 8.0 -70 Marquette 8.0 7.0 9.0 11.0 Lower Peninsula: 4-125 Emmet 11.0 8.0 11.0 -l/ -157 Presque Isle 12.0 9.0 13.0 15.0 UP combined 6.1 7.5 6.9 9.8 E 6.8 LP combined 12.2 10.8 13.0 - E 12.3 ;/ NO germination occurred. 34 Figure 2-5 graphically illustrates the relationship in germinative energy between the UP and LP at the various temperatures. At the lowest temperature of 27°C, the germ- inative energy of the birch seeds from the combined UP seedlots was 74 percent in 6 days, while for the combined LP seedlots, germination was only 14 percent for the identical length Of time. When data from temperatures of 27°, 29°, and 32°C are pooled, all the seedlots from the UP showed a much faster rate of germination than seedlots from the LP (Figure 2-6). Variation within a peninsula was small, whereas variation between the two peninsulas is readily apparent. The UP combined seedlot means reached 75 percent germination in 6.8 days, whereas the LP combined seedlot means took 12.3 days to reach 75 percent of its germination. Bxperiment III The results from this eXperiment which included all paper birch and European white birch seedlots revealed a wide pattern of genetic variation in both germination capacity and germinative energy in relation to their geo- graphic origins. Germination capacity varied both within a population and between.p0pulations. Variations in germination capacity ranged from 0.0 to 90.3 percent for all 219 seed- lots. Within an individual population (stand located in a Michigan county), germination varied widely. In many populations, single trees (seedlots) within 30 feet of each 35 Figure 2-5. Cumulative percent germination Of seven combined seedlots Of paper birch from the UP and LP of Michigan, at temperatures of 27°, 29°, 32°, and 35°C. 100 ES 29 EB 23 Cumulative GerminaIion Percent I» O 20 IO :36 —— Upper Peninsula - - - Lower Peninsula I 27°C o 29°C I 32°C a 35°C I3 14 15 16 17 37 Figure 2-6. Cumulative percent germination of all seven individual seedlots of paper birch, averaged over all three temperatures of 27°, 29°, and 32°C. 100 w c: Germination 83 Percent 29 Cumulative 33 I» O 20 ID 138 1 Days --- Upper Peninsula --—-- Lower Peninsula II 12 I I4 IS I. Ii 39 other differed by over 42 percent (4-19 and 4-20), and others within 50 yards often differed by over 63 percent (4-8 and 4-10). Therefore single trees within a population have substantial differences in their germination capacities. Substantial variation also exists between populations as evident in Table 2-7. Variations in germination capacity range from a high of 59.1 percent for population #2, located in Keweenaw county in the UP, to only 1.7 percent for pop- ulation.#l6, located in Montmorency county in the LP. Poorest germination occurred primarily in the western and north-central areas of the LP. The average percent germ- ination Of the seven pOpulations in this region was 4.8 percent. In the eastern portion of the LP the average percent germination for the five populations was much higher at 33.5 percent. Patterns in germination capacities in the UP are less conspicuous, except for an average germination capacity of 38.9 percent for all six pOpulations bordering the southern shoreline of Lake Superior in the northern portion of the UP. Table 2-7 also shows the germination capacities of the three groups of birch seed from Finland. The first group contains seed of B. verrucosa selected from a seed orchard. The average germination capacity was 41 percent, with much less variation within the group, ranging from 29.0 to 55.0 percent. As stated previously, paper birch populations from Michigan varied more than European white birch. 40 Table 2-7. Average percent germination, and the range of germination means, of 25 populations of Betula papypifera seeds from Michigan, and three groups of Finnish sources of birch, after 42 days. . Number Range Species Popufigtion Origin of Pegggnt of ' families g ' means . . - - -percent - - £219.13. W ”333%“ Upper Peninsula: 1 Ontonagon 7 35.0 17-62 2 Keweenaw 10 59.1 16-90 3 Baraga 10 32.7 6-66 4 Gogebic 7 8.7 3-13 5 Iron 10 22.4 8-44 6 Menominee 10 4.4 1-15 7 Delta 10 45.4 4-78 8 Marquette 10 22.4 2-72 9 Alger 10 37.1 13-72 10 Schoolcraft 11 13.6 1-58 11 Chippewa(W; 9 47.0 14-73 12 Chippewa(E 10 28.3 3-53 13 Mackinac 10 10. 3-28 Lower UP mean 2 . Peninsula: 14 Emmet 10 6.2 0-12 15 Antrim 5 5.3 O-lO 16 Montmorency 10~ 1.7 0- 5 17 Presque Is. 10 32.8 6-51 18 Alcona 10 16.5 0-32 19 Montcalm 4 9.2 3-17 20 Lake-Mason 5 5.1 2- 8 21 Manistee 8 2.6 l- 4 22 Missaukee 10 3.4 0- 7 23 Gladwin 9 28.7 1-61 24 Arenac 9 37.7 18-66 25 Sanilac 5 511.6. 43-61 LP mean 17.9 Overall mean: 24.1 Finnish sources: #1 Seed orchards 10 41.0 29-55 #2 Intraspecific hybrids 5 1.2 0- 6 #3 Interspecific hybrids 10 3.6 0-16 41 The second Finnish group consists of seed collected from their intraspecific provenance hybrids. Germination was very low, ranging from 0.0 to 6.0 percnet, with an aver- age of only 1.2 percent (Table 2-7). Group three consists of seed collected from their interspecific hybrids of various species of Betula with B. verrucosa. The average germination capacity was 3.6 percent, with a range of 0.0 to 16.0 percent. The hybrid of B. papypifera qu. verrucosa the highest germination capacity of the group at 16.0 percent. Thus, both the intra- and interspecific hybrids revealed low female fertility, except for the cross Ole. papyrifera x B. verrucosa. The reciprocal cross of B. verrucosa x B. papypifera was found to have surprisingly good female fertility, with 8.0 percent germination (Johnsson, 1945). Wide variations in germinative energy between the pOp- ulations of paper birch also was found (Table 2-8). Each value in the table represents the combined cumulative germ- ination of all the seedlots in that populatione Individual seedlots within a pOpulation varied in their germinative energies, but to a lesser extent than between populations. They were therefore bulked into their respective populations. Marked differences in germinative energy between the UP and LP exists (Table 2-8). Thus, 63 percent of the birch seed from the UP germinated in 7 days, whereas birch seed from the LP showed 66 percent germination only after 13 days. In addition, germination began sooner and proceeded 42 Table 2-8. Cumulative percent germination over time for Betula papyrifera and three groups of Finnish birch seed. Population Michigan Days from start of experiment No. County 2 3 5 7 13 19 Betula papypifera ------- percent ...... Upper Peninsula: 1 Ontonagon l 11 29 51 69 88 2 Keweenaw 18 43 72 8 97 99 3 Baraga 9 30 50 6 85 95 4 Gogebic g 15 20 42 71 92 5 Iron 23 48 63 87 94 6 Menominee 0 6 14 26 66 94 7 Delta 8 29 52 67 84 94 8 Marquette 10 26 31 71 91 98 9 Alger 2 l6 2 67 92 98 10 Schoolcraft 0 9 25 38 7O 89 11 Chippewa(W) 1 9 33 41 73 78 12 Chippewa(E) 3 17 48 69 88 94 13 Mackinac 0 4 13 46 82 92 Lower Peninsula: 14 Emmet O 3 18 48 83 100 15 Antrim O O 45 68 77 81 16 Montmorency 7 27 27 55 89 100 17 Presque Isle 1 13 36 56 65 91 18 Alcona O 6 32 51 73 9O 19 Montcalm O 10 20 29 49 71 20 Lake-Mason 0 0 0 11 56 75 21 Manistee 0 0 14 23 72 86 22 Missaukee 0 0 8 25 74 88 23 Gladwin O 2 9 24 55 72 24 Arenac O 1 13 25 55 75 25 Sanilac 0 ll 25 42 65 89 UP combined 6.6 22.7 46.2 63.0 84.5 93.1 LP combined 0.4 6.1 21.0 38.4 66.0 83.1 Finnish sources: #1 Seed Orchards 51 72 88 97 100 100 #2 Intraspecific hybrids O 41 41 82 100 100 #3 Interspecific hybrids 19 63 93 96 100 100 43 more rapidly in the UP, with 6.6 percent of the seeds germ- inating in the first two days, while only 0.4 percent of the seeds from the LP germinated in the same period of time. Figure 2-7 shows graphically the combined UP and LP cumulative percent germination over the 42 day period. Variations between the 25 individual populations are evident in Table 2-8. Population #2 from Keweenaw county in the UP, gave the highest rate of germination, with 72 percent of the seed germinating in just 5 days. The slowest rate of germination was for pOpulation #20, from Lake county in the LP, showing no germination in 5 days and only 11 percent after 7 days. Several patterns of variation.are apparent from Table 2-8. The 25 paper birch pOpulations are arranged in dec- reasing latitude and proceed from the western UP to eastern UP, through the northern and central LP (Figure 2—1). It is evident that the western UP populations (#1-5,7,8) as a group all began.germinating very early and proceeded at a much faster rate than the other pOpulations as a group. The eastern UP and northern-most LP populations (#9-18) as a group show intermediate results. Generally, they began germinating early but at a lesser rate than the first group, and are intermediate in germinative energy, averaging 12 percent germination in 3 days, whereas the western UP group averaged 30 percent germination after 3 days. Within this intermediate group, all ten populations show wide variations. 44 Figure 2-7. Cumulative percent germination of UP and LP combined populations over a 42 day period. 45 100 90 80 Upper Peninsula 70 I Lower Peninsula 50 Percent Germination Cumulative 10 l ' 3 6 9 12151821 24 27 30333539 42 ' Days 46 The third discernible group is the south-central LP populations (#19-25). This group shows extremely slow early germination and a much lower germinative energy, with only 4 percent germination in 3 days. At 7 days, only 29 of the seeds had germinated in this group, whereas group one and two already completed 70 and 54 percent of their total germination, respectively. The third group, and especially the southern-most populations (#19,20,23-25) have their germination extended over a very long period of time, reaching 77 percent of their total germination after 19 days; Also included in Table 2-8 is the cumulative germ— ination percents of.the Finnish seed and hybrids. The single trees that comprise the seed orchard group initiated germination very early, and proceeded at a very fast rate. Twenty-four hours after the beginning of the experiment, 2 percent of the seeds already germinated. By the second day, 51 percent of the seeds had germinated, showing the maximum germinative energy of both species of birch. It should be understood that EurOpean white birch seed is from selected parents and not comprised of wild pOpulations as paper birch. Groups two and three had very few seed germinate, and the results are similar to some populations of paper birch. One important difference between the two species of birch is that B. verrucosa had no germination past the 13th day, unlike B. a ifera, which required at least 42 days in a number of seedlots. 47 Discussion In addition to determining the Optimum temperatures for the germination of paper birch seed, the present study also demonstrated significant differences in germination capacity and germinative energy in response to temperature and geographic origins. The response of birch germination to temperature varies greatly among different seedlots. Generally, as temperature is increased germination capacity and energy is increased. Seedlots from the UP all showed a much wider optimum temp- erature range as well as increased germinative energy in comparison with seedlots from the LP. The optimum temperature range for paper birch seed germination percentage and rate is between 27°— 32°C. Germination occurred from 10° to 43°C in the UP seed, revealing extreme plasticity, whereas the LP seedlots showed much reduced germination at 10° and 35°C with 43°C being lethal. The UP seedlots had an extremely wide range of temperatures Optimum for germination (21°-32°C), with no significant differences existing between 210 to 32°C, whereas the LP seedlots had a much narrower range of optimum temperatures (27°-29°C). These findings are contradictory to those reported by Govorukha and Manaev (1971) for B. pppggpgpg and B. xgpppppgg, in which seeds from more southern latitudes had a wider range of Optimum temperatures, and northern origins required higher temperatures. The work reported by Fraser (1971) on 48 white spruce provenances in Canada is in general agreement with these findings, with northern provenances germinating at lower temperatures. There are several possible reasons for the wide range of germination temperatures in the UP seed. Natural shed- ding of seeds takes place from late fall to early spring. During the winter months, the seed remains in an environment unfavorable for germination. In the UP, seed dispersal coincides with the arrival of consistently cooler temp- eratures, whereas in the LP, dispersed seed is subjected to wide fluctuations in temperatures. Sources in the LP may have adapted to safeguard the species from germinating in the late-fall when temperatures may reach 10°C or greater for extended periods of time. These warm periods are often followed by cold periods severe enough to kill any newly emerged paper birch seedlings. It is possible therefore that ecotypes producing seeds that germinate with a wide range of temperatures allowing germination would tend to be eliminated to a greater extent in the southern LP than in the northern latitudes of the UP. Those seed produced in the UP would be subjected to less deviation in the fall temperatures, thereby allowing a wider range of germination temperatures. A similar situation exists with regard to spring temp- eratures. The LP frequently experiences late winter thaws where temperatures reach 10°C for extended periods of time, followed by a return to freezing temperatures that would 49 kill any newly emerged seed. This would be less likely to occur in the UP where the growing seasons are short and where frozen soils, protective during the winter, warm up slowly. Another possibility may be a requirement of LP seedlots to accumulate a certain "heat sum" before triggering germ- ination. Thus, at 10°C, the LP seedlots may require a large number of degree-hours, before reaching the triggering "heat sum" or the temperature may be too low to even begin accumulating degree-hours. Phillips (1922) noted that fruit trees apparently required more heat before blooming in the South than in the North. Boyer (1972) found that longleaf pine from North Carolina required fewer degree-hours for peak pollen shed- ding, than those in southwestern Alabama. Experiments I and II also revealed differences in the initiation phase of seed germination. At temperatures ranging from 10°-29°C, LP seedlots did not initiate germ- ination until 48 hours after the UP seedlots had already begun germination. At temperatures of 32°C, germination was initiated on the same day for all seedlots from both penin- sulas. Rapidity of germination can have an important bearing upon seedling survival and establishment. Bxpgpppggp§_;_gpg 2; revealed significant differences in germinative energy due to seedlot origin and temperature. Generally, as temperature was increased, germinative energy was increased. 50 High and consistent germinative energy for seedlots in the UP occurred from 27°-32°C. The optimum temperature for the LP seedlots ranged from 27°-32°C, with 29°C providing the highest germination rate (Figure 2-5). The two experim- ents revealed that, at all temperatures, UP seedlots began germination sooner, proceeded more rapidly, and reached optimum levels sooner than LP seedlots. Experiment III was designed to elucidate the specific pattern of geographic variation found in the previous two eXperiments, but on a much larger scale. The results confirm the pattern of variation found in the previous two experiments. Variation exists within a pOpulation, between populations, and between upper and lower Michigan in both germination capacity and germinative energy. Germination between all seedlots ranged from 0.0 to 90.3 percent. Within an individual population, encompassing about a hectare in size, differences ranged from 16 to 90 percent. Significant differences between population in germinative energy occurred, varying by a maximum of 54.7 percent. One population in Presque Isle county (#17), for example, averaged 32.8 percent germination while another population 40 miles away averaged only 1.7 percent. Variability in germination capacity can be due to a number of factors. Seed viability depends greatly on local weather conditions during pollination, fertilization, and subsequent seed development. Quality of seed varies from year to year, and is usually good in years with a 51 heavy seed crop, but poor when the seed crop is light (marquis, 1969). A study of germination capacity of paper birch seed in Maine through 1958-1960 revealed germination of 77 Percent during a heavy seed year, versus germination of 13 and 24 percent during the two previously poorer seed years (Bjorkbom et al., 1965). Results here indicate that size of the seed crOp of the parent tree has little influence on germination capacity. Each parent tree was assigned a number relating to the size of the seed crOp: l=1ight, 2=medium, and 3=heavy. Cor- relation analysis was used on all 219 seedlots to determine if a relationship existed between the seed crOp of the parent trees with percent germination of the seedlots. A non-significant correlation (r) of 0.06 was found. In many cases trees with a very light seed crop had very high germ- ination capacities. Within a population, seed crOps varied from light to very heavy. In some populations the ten or so single trees varied only between the ranges of light- medium, or medium-heavy. No single population had all single trees with the same seed crop rating. For the 25 populations, correlation between average seed crop and per- cent germination also were non-significant (r=0.03) Germinative energy, or rate of germination, revealed wide variations, with several observable patterns. The combined UP populations reached 63 percent germination in 7 days, whereas the combined LP populations took nearly 13 days to reach the same level. The most distinctive 52 difference is between the western UP and the south-central LP populations. When grouping populations in these larger areas, the western UP group revealed very high germinative energy, beginning germination very early and prodeeding at a much faster rate than the south-central LP populations. For example, 70 percent of the birch seed frOm the western UP germinated in 7 days, whereas birch seed from the south- central LP showed only 29 percent in the same time period. A large amount of variation also existed between the 25 individual populations of paper birch. Bjorkbom (1971), studying paper birch seed germination in Maine, reported that germinative energy reached higher levels in the better seed years, with germination beginning sooner and proceeding more rapidly than in less productive seed years. Results reported here tend to disagree with the above findings, although this analysis was on seeds collected in a single seed year and on a population basis, each comprised of individual trees. Correlation analysis was performed on the 25 populations of paper birch to relate the average parental seed crop of the population to cumulative percent germination occurring after 5 days. No correlation was found (r=—0.09), indicating that germinative energy is not related to size of the paper birch parental seed crOp. Since germinative energy expresses how rapidly germ- ination proceeds, it is better than germination capacity as a practical measure of the average viability and quality 53 of a seed crop. For artificial regeneration, high germ- inative energy is an important requirement. The practical advantage to a greenhouse and nursery manager is the ability of the seedlings to get off to a prompt and even start, producing uniform growing stock responsive to favorable growing conditions. Ecologically, rapidity of germination has an important bearing upon early seedling survival and subsequent estab- lishment. The more rapid rate of germination for the UP seedlots may have a considerable survival value for these northern seedlots, i.e., their high rate of germination allows the fragile seedlings to quickly take advantage of warm growing days in the spring and therefore make better use of the shorter growing season. In addition, seedlings will have a maximum length of time to become established before subjected to summer drought and high temperatures. During this investigation of the germination behavior of paper birch seed, a few seed showed abnormal germination. In studying 10,000 yellow birch seed, Maini and Wang (1967) reported the occurrence of "twin seedlings" at a frequency of 0.12 percent, and in a few other seedlings the cotyledons emerged from the pericarp before the radicle. Of the approximately 34,000 seeds handled in these experiments, the emergence of "twin.seedlings" was encountered in four seeds, giving a frequency of 0.012 percent. About 10-15 seeds showed reversed germination or early cotyledon emergence. 54 Summary and Conpiusions This study appears to have revealed some important variation patterns in the germination of paper birch seeds. Temperature and geographic origin each had a significant effect on germination capacity and germinative energy. Seedlots from the UP of Michigan showed a much wider range of temperatures Optimum for germination, as well as increased germinative energy, in comparison with LP seedlots of Mich- igan. Distinct differences were also observed in the initiation of seed germination, with the UP seedlots begin- ning first. Germination capacity and germinative energy were not correlated with the size of the parental seed crop. Genetic variation exists both between and within pop- ulations of paper birch in germination capacity and germ- inative energy. POpulations in the western UP were charact- erized by germination beginning much sooner, proceeding more rapidly, and reaching optimum levels earlier than populations located in the south-central LP. If these observations made in a controlled environment are indicative of the actual responses occurring in natural environments, UP and LP seedlots each have adapted to their local environments, providing continual renewal of the species. European white birch initiates seed germination very early, and proceeds at a much faster rate than the paper birch seedlots. In addition, European white birch ended germination after 13 days, whereas paper birch extended germ- ination for a period of 42 days. CHAPTER III GEOGRAPHIC VARIATION IN THE GROWTH RATES OF SEEDLINGS Introduction Variation within a species creates Opportunities for locating superior types. Without variation there would be nothing to improve upon. Species with a very large natural range, such as paper birch, contain considerable genetic variation and diversity. Variation is the result of the interaction of two forces: first and most important is the genetic make-up of the individuals, and secondly, the influence of the local environment. This variation may be either clinal, as shown by gradual transitions following environmental changes, or ecOtypical and composed of fairly distinct subdivisions or ecotypes. To determine the presence and amount of geographic variation in any species, large numbers of seedlings from a range of geographic regions need to be tested under similar conditions and distributed throughout its range over a variety of climatic and edaphic conditions. A prerequisite to the improvement of a particular species is the need to know how much genetic variation is present. While this information is known for many import- 55 56 ant commercial trees such as eastern white pine, jack pine, red pine, white spruce, Douglas—fir, etc., research on paper birch has been scanty. Geographic origin studies on paper birch have been absent. The past geographic variation studies in birch has been limited mainly to Betula alle haniensis, yellow birch. Clausen (1968a) found that catkin and fruit character- istics of yellow birch vary in an apparently random manner over its natural range and differ among individual trees within the stands. Morphological variation in characters such as leaf size and shape, catkins, and seed is believed due to natural hybridization and introgression (Clausen, 1962: Dugle, 1966: Dancik and Barnes, 1972). In some species of birch this variation is believed to be due in part to differences in the chromosome numbers. In paper birch, Brittain and Grant (1966, 1967) have shown that chromosome counts of 56, 70, and 84 with 56 and 70 most prevalent in eastern Canada, and 84 more common in northwest Canada. In terms of growth characteristics, wide variations has been demonstrated for yellow birch. Clausen (1968b) reported that second-year nursery height growth of 55 seed sources of yellow birch, comprising approximately 10 trees per stand, varied greatly but essentially in a random manner. Height was not correlated with latitude, longitude, growing season, average January or July temperatures, or annual precipitation. In contrast to the random variation 57 in height growth, the study demonstrated that yellow birch exhibits clinal variation in growth cessation. Clausen and Garrett (1969) reported the one-year results from a 199 single tree progeny test to determine whether the large height variability observed in the 55 seed-source study might be due to the fact that the seed- lots used were a mixture from an average of 10 trees per stand. After the first growing season , seedling height was significantly different not only among the 21 sources but also among the trees within each source. The results reported after 4 years showed that certain provenances were much more variable than others, and the amount of variation in a particular provenance frequently differed from year-to-year. Within-provenance variation in 4-year heights of yellow birch seedlings was often greater than differences among the 21 widely separated provenances. It was concluded that due to the large amount of individual tree variation present in yellow birch a selection program should focus on single tree selection (Clausen, 1972). Because of the growing importance of paper birch as a timber producing tree and potential for artificial reg- eneration, information on geographic variation in growth rate, along with other desirable traits, is an improtant first step towards its improvement. Due to the apparent ample geographic variation and the large amount of individual tree variation in yellow /— 58 birch, the same possibilities may exist in paper birch. Therefore, this initial study was undertaken to investigate the genetic variation patterns present in this species. Materials and Methoq§ The seed used for these studies originated from the state-wide collections of fruiting catkins during September of 1975. Details of the collection procedure and locations were described in Chapter II. This study consisted of three separate experiments: Bxperiment I included all 219 seedlots (single half-sib families) of B. papyriiera and 25 seedlots of Finnish birch. Bxpgpiment II consists of 42 seedlots (populations) of paper birch distributed over its natural range. Bxperiment III consisted of essentially a re-analysis, in part, of the previous two eXperiments. Included was a total of 74 seedlots, of which 48 were single families of paper birch from the Michigan state-wide test, 11 single families re-collected in September of 1976, 9 single families from the rangewide test, 4 seedlots of B. verrucosa, and two miscellaneous species of birch. Experiment I This initial experiment on variations in juvenile height growth of the seedling progeny contained all 219 single families, representing 25 distinct populations, collected in Michigan in Fall of 1975. Exact location of the 25 populations are shown in Figure 3-1. In addition 59 Figure 3-1. Location and population numbers of the 25 paper birch populations used in Experiment I. 61 25 seedlots of birch seed from Finland, consisting of three groups, as discussed in Cha ter II, are included in the study. On January 6, 1976 all 219 seedlots of paper birch and the 25 seedlots of European white birch were sown in a greenhouse progeny test at Michigan State University's Tree Research Center. The seeds were sown in polycoated (Zeiset) paper plant bands 5cm x 5cm x 28cm deep, containing a soil mixture of peat-vermiculite (Redi-Earth, manfactured by Terra Lite). The paper plant bands were packed in milk cases, each containing 36 bands. The number of seeds sowed ranged from 3 to 150 depending on the percent germinat- ion of that seedlot. The seeds were not covered but were watered frequently during the germination phase. Greenhouse temperatures during germination was inc- reased to approximately 27°C to promote rapid and uniform germination. Transplanting and thinning of the seedlings took place one month after sowing when the seedlings were 0.5- 1.0cm in height. At this time they were thinned down to the three most vigorous seedlings. After two more weeks they were finally thinned to the most vigorous seedling in each band. Fertilization began with Peters 20-20-20, two weeks after sowing, and was applied with a hozon injector (1231) throughout the observation period. A fungicide (Benlate) was applied to control fungus. No damping-off was observed. 62 Continuous light was provided by high-output (H0), cool white, 8 foot fluorescent lights which were placed 7 feet above the cases. The lights provided approximately 120 ft-candles at the soil surface, and were operated during the night. Temperatures in the greenhouse was maintained at approximately the 23-27°C range, both day and night through- out the 4-month growth period. The experimental design consists of a randomized complete block with 5 replications. Each replicate contained one row of six seedlings per seedlot, for 234 seedlots. Each replication contained 39 cases, of which each case contained 6 randomly chosen seedlots, with each seedlot being represented by a single row of six seedlings, giving a total of 36 trees per case. In May 1976, 4 months after sowing, the containerized seedlings were measured and moved outside to the nursery to condition them for field planting. Analysis of variance was made on growth rate of the seedlings at age 4-months, prior to their removal from the greenhouse. The LSD test was employed to determine the significance of height differences between seedlots, pop- ulations, and peninsulas. The tallest three trees per plot were measured and seedlot means were used as items. Experiment II The objectives of this experiment on the variations in juvenile height growth of paper birch seedlings is identical 63 to the previous experiment. The only major difference is the materials used are not part of the state-wide collection made in 1975. but instead include seed from over the entire range of paper birch. In the Fall of 1975, David S. Canavera, Assistant Professor, University of Maine, initiated a rangewide col- lection of various seed origins of paper birch through the cooperation of various private and governmental agencies. The seed used here is a portion of those seed sources. Seeds from a given locality were collected from each of about ten average trees per stand. Catkins collected within a local area were separated into two groups. Some seedlots were identified by ten single tree seedlots and others were bulked into one seedlot comprising ten individual trees. Seedlots used in this study are comprised of the latter, a bulking of about ten trees per stand. Seedlots were accompanied by data on latitude, longitude, elevation, and collection date. Exact locations of the 42 seedlots used in this study are shown in Figure 3-2. The origins and pertinent climatic data are presented in Table 3-5. On March 24, 1976 the 42 seedlots (populations) from throughout the range of paper birch were sown in a green- house at M.S.U.'s Tree Research Center. Details of the method of sowing, handling, and growing conditions, are identical to those described in this Chapter under EXperiment I. 64 Figure 3-2. Location and pOpulation numbers of the 42 paper birch populations used in Experiment II. 65 £83 .39!— uo can! can. \\ :qflifld...” villi: / I a. .. (I \N‘. iis. J! i.....ii._...._.i. ... . ‘1‘. . \ . I. . . . x... . . a (L / 7 x ._ . ... .z u. . . . ..... . . . ' .\|. FI.‘ . \ . f0 II. . a, o. D. . C I O m w IL! I .. NJWWMWI IIL. I IIIIL f\ %.@L~ \ I o .I .. .. r - / \J... I‘m/.I. . / . : 4.0 .I.” O . .~. . . nu... . _ m .c . . II . .II .. .N‘llv J58. ' In. ..I. O \ I: H \\ . . a 1 . e .H. 1 .. , . . .... as... . 1 ) .o. . \ \. ., . \\ ‘ .INII. . \ . .a . . . . . e 2.. . . A» I I .u. .I a 1.. x , .. I.’ .151. .s . I. . b L . _. .I ~ . .(f if. 0‘ ... . flirty. . TI . .f , . \\ ._. .. .‘Arvr , a .v. I9 KHz/[WI an, I I 91 < I. . I z .\ Id . a a, .1 . . .. “M n ,o/ .. , /V\Vé O . .. x x. / .. e. z .p / . I s \.I .o. I J. a L a. t 5... tom -3. \nw 66 Experimental design consists of a randomized complete block with 6 replications. In August 1976, 5-months after sowing, the containerized seedlings were measured and moved outside to the nursery to condition them for field planting. Analysis of variance was made on growth rates of the seedlings at age 5-months, prior to removal from the greenr house. Measurements were made on the tallest three trees in each 6-tree plot. LSD was used to determine the sign- ificance of height differences between the seedlots. Seedlots means were used as items. Simple correlations, using seedlot means as items, were calculated between latitude, longitude, and length of the growing season at place of seed origin, and juvenile height growth of 5-month old seedlings. Experiment III This final study was designed to measure variation in the juvenile height growth rates of seedlings. The major difference is that this is essentially a re-analysis, in part, of the two previous experiments. The objective is to compare the results observed in the previous two exp- eriments. Materials used in this study are a part of those previously used, but not all the seedlots were used for both. This progeny test contained a total of 74 seedlots of which 48 were single tree families, contained in 8 67 Michigan pOpulations, collected in 1975. In addition, 11 single families originally collected in 1975 were re-col- lected in 1976. This means that 11 seedlots out of the 48 original seedlots have both 1975 and 1976 seed represented. Also included are 9 single tree families from the range- wide test collected from three geographic areas; 4 seedlots of Finnish origin: and two miscellaneous species of birch. The exact location of the 8 Michigan populations are rep- resented in Figure 3-3. The seedlots for this experiment were selected to cover a wide range of climatic conditions and were based upon previous results of the juvenile height growth data. Two populations were selected in each climatic area of Michigan: the western UP, eastern UP, northern LP, and southern LP. Within each area, the fastest and slowest populations were selected on the basis of the results of Experiment I. Within each pOpulation, the three fastest and the three slowest individual families were selected. Therefore, there were 8 pOpulations, 6 families per pOp- ulation, and a total of 48 seedlots. The 11 re-collected families are from three LP populations, located in three counties: Missaukee, Gladwin, and Arenac. In Ex eriment II, seedlots included a bulked sample of ten average trees per stand. In this eXperiment, nine individual families were randomly chosen, three families from each of three widely separated areas: Montana, Quebec, and Alberta. 68 Figure 3-3. Location and population numbers of the 8 paper birch pOpulations from Michigan used in Experiment III. Populations marked with a asterisk are represented by seed collected in 1975 and 1976. 69 . alum an MOI“ lumu ”IO!“ [III m7 ION maul mam spun an! (AIM < I“ 1W”? .JKM ”I‘M' ——‘ um m ‘1! I“! l I“! 70 The four seedlots of Finnish origin represented wide areas of Finland. Three are B. verrucosa individual tree families selected from a seed orchard, and the other is a B. verrucosa interspecific hybrid. On October 20, 1976 all 74 seedlots were sown in a greenhouse progeny test at M.S.U.'s Tree Research Center. The seeds were sown in polycoated (Zeiset) paper plant bands 7.5cm x 7.50m x 28cm deep, containing a soil mixture of peat, perlite, and vermiculite (1:1:1). The plant bands were packed in milk cases, each containing 16 bands. Seed was sown in prOportion to the germination percentage of the seedlot. Fertilization with Peters 20-19-18 (Peat-lite special) began two weeks after sowing and was applied with a hozon injector (12:1) throughout the observation period. Benlate was applied periodically to inhibit fungus growth. Supplimental light were used as described in Experiment I. The experimental design consisted of randomized complete blocks with four replications. Each replication contained one row of four seedlings per seedlot (family), randomly distributed, for all 74 seedlots. On.March l, 1977, 4-months after sowing, the final of the four periodic measurements were made. Three weeks later, the containerized seedlings were cut-off 3-4" from the root collar. Re-sprouting ability was periodically observed. Analysis of variance was made on growth rates of the seedlings at ages 2%, 3, 3%, and 4-months. LSD was used 71 to determine the significance of height differences between families, pOpulations, families within-populations, and between peninsulas. Seedlot means were used as items. Simple correlations, using family and population means as items, were calculated between 4-month juvenile height growth and latitude, longitude, length of the growing season, 2%- and 3-month height growth. Results and Discussion Under the conditions of greenhouse culture and 24-hours continuous light (long-days), variation in seedling height growth and leaf morphology were observed. It should be noted that these results are observed on seedlings subjected to an unnatural artificial environment of continuous light, and controlled temperatures. Early evaluations of the results reported here should be correlated with the perform- ance of similar materials grown under natural field con- ditions. Juvenile-mature correlations of future height growth of these seedlings, established in future plant- ations, will give an indication of the reliability and usefulness of such early observations in a genetic improve- ment program. Experiment I Due to cultural difficulties encountered during the initial growth period of the seedlings, statistical analysis was impossible. These cultural problems origin- 72 ated from both the variability of the soil mix and watering. A definite "case effect" was Observed, with some cases showing very slow growth common to all six seedlots in that case. The reasons are thought to be due to uneven packing of the soil mix (Redi-Earth) and overwatering. If a case had been packed too tightly, the soil water-holding capacity was greatly increased and waterlogging occurred. Results observed indicate that this phenomenon occurs as a result of the physical nature of the soil mix itself. Due to the initial dry form and the 280m depth of the soil column, wetting was difficult, and once wet, drying was impossible and waterlogging occurred. The present study was undertaken to determine how much of the variation in juvenile growth rates can.be ascribed to differences both within- and between-populations of paper birch. Results obtained under these growing conditions reveal that within a population (stand) of paper birch, large variation exists in juvenile height growth (Table 3-1). Certain pOpulations were much more variable in height growth than others, but the within-population.(family) variation was greater than differences among pOpulations. The average 4-month height of the study was 59.20m, with a range between the two extreme pOpulations of just 19 percent. Only one individual population, in Antrim county population (#15), with a range of 10 percent of its mean, had less within-pOpulation (family) variation than 73 Table 3-1. Four-month relative heights of 25 paper birch seed sources ex ressed as percent of the overall mean (Experimen I). Michigan Population Michigan Number 4-month Range Of . . of relative family Pen1nsula NO county.families heights means A/ - - - -percent - - - - Upper: 1 Ontonagon 7 96 84-114 2 Keweenaw 10 105 79-119 3 Baraga 10 108 80-116 4 Gogebic 7 98 83-119 5 Iron 10 108 84-120 6 Menominee 10 100 88-108 7 Delta 10 108 92-117 8 Marquette 10 96 88-123 9 Al er 10 98 86-116 10 So oolcraft 11 101 92-112 11 Chippewa(w) 9 95 86-121 12 Chippewa(E) 10 100 90-115 13 Mackinac 10 98 84-109 Lower: 14 Emmet 10 95 84-116 15 Antrim 5 93 95-105 16 Montmorency 10 98 83-116 17 Presque Is. 10 96 82-114 18 Alcona 10 101 88-113 19 Montcalm' 4 96 86—114 20 Lake-Mason 5 105 84-113 21 Manistee 8 91 87-128 22 Missaukee 10 106 86-108 23 Gladwin 9 103 84-113 24 Arenac 9 100 85-115 25 Sanilac 5 98 81-116 219 Overall mean: 59.2cm Upper Peninsula, combined 101 Lower Peninsula, combined 98 ;/ Expressed as percent of population means. 74 the overall difference among the populations of 19 percent. All other populations had a greater amount of within-pop- ulation variation, with the maximum of 41 percent of the mean for the pOpulation in Manistee county (#21) in north- west lower Michigan (Table 3-1). Maximum individual family variation in height growth ranged from a high of 77cm (130% of the overall mean) for family (35) located in Iron county, to 47cm (79% of the overall mean) for family (129) located in Emmet county in the LP. Thus, the Iron county family grew 61 percent as much as the Emmet county family, which was poorest. Population differences after combining the families within each population are small. Two populations, Baraga: and Manistee, separated by a range gap of about four miles between the two peninsulas of Michigan, and a distance of over 300 miles, differed only by 19 percent. Other pop- ulations with the same distribution had nearly identical 4-month heights. When family means for the UP and LP of Michigan were combined, the differences between the two peninsulas was only 1.6cm (Table 3-1). Family evaluation after only 4-months would be highly speculative. However, 4-month results indicate family differences in juvenile height growth are substantial. Sixty-six families that were present in all 5 replications were averaged to locate the six best and six poorest families (Table 3-2). Averaged across all 5 replicates, 75 Table 3-2. Four-month height of the six best and six poor- est families of all 66 families represented in all five replicates. Six Seedlot Michigan 4-month Mean families No. County helghts _____ cm _ - - - _ Best: 64 Delta 75 18 Baraga 74 34 Gogebic 69 42 Iron 69 62 Delta 68 98 Chippewa (W) 68 70.6 Poorest: 185 Manistee 47 31 Gogebic 48 78 Alger 5O 65 Marquette 51 66 Marquette 51 184 Manistee 51 49.7 Mean of the six best families: 70.6 Overall mean of all 66 families: 59.5 Mean of the six poorest families: 49.7 Percent improvement over the overall mean: 16.0% 76 the six best families averaged 70.6cm in total height after 4-months of accelerated growth. The six poorest families averaged 49.7cm.in total height. An improvement of 16 percent is obtained when the six best families are compared to the overall mean of the 66 families. Results also revealed wide variations in the juvenile growth rates of the 25 seedlots of European white birch included in the progeny test of paper birch. The 25 seed- lots of birch seed, consisting of three separate groups, also showed wide variation in growth rates under these conditions. Maximum individual seedlot variation in height growth ranged for 860m (145% of the overall mean) for seedlot number 1459 located in southern Finland, to 44cm (74% of the overall mean) for seedlot number 1456 also located in south- ern Finland, a difference of 95 percent. Taken as a group, the ten interspecific hybrids performed the best, averaging 110 percent of the mean. The 'B. papypifera x‘B. verrucosa hybrid averaged 68cm in total height, 15 percent over the overall study mean of paper birch, and 9 percent over the mean for EurOpean white birch. Due to the diversity of the cases in terms of moisture, a comparative analysis of the two birches is difficult. It was readily Observed however, that when the European birch occurred in a Optimum case, it outgrew the paper birch seedlings. After only 4-months continuous growth, some individual progeny of both birches reached 115cm in height. 77 Expgpiment II Seedlots included in this experiment consisted of seed collected from about ten individual trees per stand, bulked together to make a seedlot. The seedlots used in Experiment 7;, differ from those used here which are single tree fam- ilies and not pOpulations. Cultural problems encountered in Experiment I were virtually eliminated, and statistical analysis was there- fore possible. Results obtained under these growing conditions revealed wide variation in juvenile height growth of the 42 populations of paper birch distributed over its natural range. The average 5-month height of the populations was 81.6cm, with a range between the two extreme populations, Minnesota (251) at 98cm (120% of the overall mean) and British Columbia (265) at 620m (76% of the overall mean), of 36cm, or 58 percent taller than the poorest pOpulation (Table 3-3). For comparison, Clausen (1969) reported for 55 seed- sources of yellow birch, the best provenances was 117 percent ’taller than the poorest provenance after three years in the nursery, averaging 47.8cm. Analysis of variance revealed that variations in the growth rates of seedlings of paper birch, grown under accelerated conditions, to be significant (Table 3-4). Differences between pOpulations were significant at the 5-percent level and highly significant differences ’78 Table 3-3. Five-month relative height growth and origin data of 42 paper birch populations, expressed as percent of the overall mean. 000- Pop State Degpees of- Grow- No. of 5-month Relative graphic No. or N. W. Elev. m- relative heights of Area ° Province Lat. Long. Season-/ ilies-/ heights geog. area -ft- -days- - - - percent - - - N.W. U.S. 248 Alaska 59.2 135.4 250 122 4 82 and 265 Brit. Col. 3.7 120.8 2200 100 10 76 Canada: 292 Brit. Col. 9.3 121.5 100 200 10 108 247 Montana 48.0 114.0 3150 149 10 99 266 Montana 47.? 113.8 3800 138 10 103 4 9 Western 261 Alberta 55.2 114.6 2200 90 10 99 Plains: 262 Alberta 55.0 114.6 1750 85 10 98 285 Alberta 54.1 115.8 3350 90 2 98 286 Alberta 54.1 115.8 3350 90 5 104 287 Alberta 54.1 115.8 3350 90 7 103 ' 100 Central 281 Saskatch. 52.0 105.0 1500 70 — 114 Plains: 282 Saskatch. 52.0 105.0 1700 80 - 94 283 Saskatch. 53.5 106.0 1950 70 - 113 107 Northern 269 Ontario 49.5 93.0 1200 108 10 98 Lake 267 Ontario 49.9 92.3 1300 100 10 114 States 273 Ontario 49.8 86.7 1100 75 10 113 275 Ontario 48.3 89.2 620 95 10 115 110 Western 251 Minnesota 47.5 93.0 1450 111 10 120 Lake 252 Minnesota 44.9 93.6 945 139 10 115 States 253 Minnesota 44.0 91.5 1250 158 12 71 254 Wisconsin 45.5 89.5 - 127 10 107 256 Wisconsin 43.2 90 2 850 146 9 92 260 Wisconsin 44.8 88.5 - 126 10 93 263 Wisconsin 44.5 89.5 1100 130 7 118 289 Wisconsin 47.0 91.0 830 116 10 96 101 N.E. Lake 271 Ontario 46.3 83.4 700 110 10 102 States 274 Ontario 46.7 80.6 - 100 10 113 280 Ontario 48.1 83.5 1400 80 10 78 268 Ontario 47.1 79.8 950 90 10 99 8 9 3.3. Lake 277 Ontario 44.8 81.1 590 145 10 33 States 255 Ontario 44.3 80.0 700 140 12 91 272 Ontario 45.0 80.3 600 140 5 91 88 N.E. New 279 Ontario 45.5 77.0 - 125 10 104 England 259 Quebec 46.9 76.1 1100 175 10 92 and 258 Quebec 47.0 71. 700 175 7 96 Canada 227 Quebec 49.2 66. 1200 150 6 102 2 Vermont 44.5 72.9 900 95 10 118 , 102 Eastern 264 Mass. 42.7 73.2 1760 144 10 107 U.S. 291 Mass. 42.6 73.0 1980 156 10 107 250 Conn. 41.9 73. 3 1300 131 6 98 284 Pennsyl. 40.8 78.01500 150 6 86 290 Pennsyl. 41.3 75.51600 160 3 91 98 Overall mean 81.60m LSD (0.05) of a population mean 25.7 percent ;/ Average number of days without a killing frost. 2/ Number of individual trees collected per population. 79 Table 3-4. Analysis of variance of juvenile height growthl/ of 42 paper birch seedlots 5-months from seed. Source of variation d.f. M.S. F-value Between blocks 5 643.03 1.86** Between seedlots (Pops.) 41 559.18 1.62* Between geographic areas 8 415.47 0.70 Between seedlots w/areas 33 594.10 1.72** Error 205 345-38 - ;/ The analysis is based on plot means. Tallest three trees per plot. * Significant at the 5-percent level. ** Significant at the l-percent level. existed between the geographic areas. Correlation analyses showed that height growth of 5- month-old seedlings grown under accelerated conditions is not correlated with latitude or longitude, and appears to vary in a random manner (Table 3-5). Table 3-5. Correlation between juvenile height growth of 42 paper birch seedlots and environmental variables of the seedlots. Environmental variable Height growth r value Latitude 0.01 Longitude -0.07 . . *1/ Length of growxng season -O.4l * Significant at the 5-percent level. 1/ Based on 24 origins with reliable data on the length of the growing season. Includes all origins from Minnesota, Wisconsin, Quebec, Ontario, and Vermont. 80 Correlation analysis revealed a weak correlation of 5-month height with length of the growing season. Of interest is the fact that the correlation coefficient is negative, meaning that under conditions of continuous light those populations from areas with a shorter growing season tended to grow the fastest. The same negative trend was also apparent in EXperiment I, with a non-significant r=-O.39. A possible explanation for this effect could be due to the fact that under conditions of continuous light, populations in areas with a short growing season take full advantage of the season's length, and subsequently grow at a faster rate. Likewise, pOpulations in areas with a longer growing season have evolved mechanisms to extend the growth phase through a longer period of time, and there- fore, not be as responsive to continuous light. The data not only revealed differences in juvenile height growth but also in leaf morphology, leaf number, branching, and stem pubescence. Variation in stem pubescence ranged from densely pubescent to a sticky glandular exudate. All the seedlings from the Saskatchewan and Alberta provinces of Canada were absent of any stem pubescence and produced a glandular exudate, which appeared warty in nature. These warts appear as white flecks or scabs on the stems. Brittain and Grant (1975) reported similar observations on young paper birch seedlings collected from Prince Edward Island and the 81 Rocky Mountain area of Alberta. In addition, all seedlings from the Sasketchewan and Alberta area produced a large number of leaves and a prolific branching habit. In terms of growth rate all the seedlings from these areas ranked equal to or above the overall mean (Table 3-3). Seedlings from these areas initiated growth earlier in the spring than any other seed- lots of paper birch when observed the following year. It is possible due to its overlapping range with a small tree or shrub, B. fontinalis, that these trees are natural hybrids. Dugle (1966) reports these two species sometimes hybridize, which may account for the warty condition of the seedlings, common on the twigs of western red birch (B. fontinalis) Wartiness also appeared in a more random manner on two widely separated populations. POpulation (258), located in Quebec exhibited the condition. In addition, population (249) from Vermont produced about one-half the seedlings pubescent and the other half glandular, revealing additional variation within a particular stand of ten trees. Additional morphological variation was observed to be associated within individual populations. An Ontario pop- ulation (280) exhibited an unusual pattern of leaf venation, shape, and serrations on all seedlings. Population (275), located on the north shore of Lake Superior in Ontario characteristically produced larger leaves than other pop- ulations of paper birch. 82 It is revelant to point out that the populations used are distributed over a wide geographic area. They are comprised of individuals from varying age groups, diverse sites, and a wide range of climatic conditions. No tax- onomic attempt was made to measure and catagorize the morphological variability present. Instead, some of the more observable characters were mentioned. Obviously con- siderable morphological variability exists between different populations of paper birch, which in part reflects its wide geographic range, overlapping with many other species of birch, and the diverse ecological sites occupied by this species. Chapter IV examines in more details the developmental and morphological variability observed in certain characters of seedlings. Experiment III Unlike the two previous experiments, this experiment is basically a re-analysis of the results observed in the previous two experiments on juvenile height growth of paper birch seedlings. Cultural problems encountered in the past experiments were eliminated entirely, and statistical analysis was possible. Only relative comparisons will be given between this experiment and the previous two. Major differences included a different soil mixture, larger plant bands (7.50m2 x 280m 2 versus 50m x 28cm), the experiment was initiated at a 83 different time of the year, and there was very good control of watering and cultural conditions. Results obtained under these growing conditions again revealed wide variation in juvenile height growth as reported in Experiment I. Table 3-6 gives the relative growth rates of the paper birch and EurOpean white birch populations used in this study. The material consisted of 8 pOpulations from Michigan, collected in 1975. with each population represented by six individual families. In addition, 3 populations re-collected in 1976 are included. Also three widely separated populations from the rangewide study, each con- taining three individual families. Seedlots of Finnish origin are included, represented by B. verrucosa seed orchard material (3 seedlots) and one interspecific hybrids. Two miscellaneous species of birch are included, thereby giving a total of 74 seedlots in the study. Analysis of the results are presented in two ways: the first, analysis contains only the 48 single families collected in Michigan in 1975, and the second analysis included all 74 seedlots. Analysis of the 48 individual families from Michigan at 4-months show that certain pOpulations were much more variable in height growth than others, and within-pop- ulation (family) variation was greater than differences among populations (Table 3-6). 84 Table 3-6. Species, origin, and 4-month relative height growth expressed as percent of the overall mean for 8 populations of paper birch collected in 1975, 3 pOpulations re-collected in 1976, three range-wide populations, and two other species of birch (EXperiment III). Number 4~month Range of Species Popufigtion Origin of relative family 1/ ' families heights means Betula papyrifera ‘ ' ’ percent ' ' Michigan 1975 Collections: County 3 Baraga 6 107b 2/ 82-121 4 Gogebic 6 95bc 80-111 7 Delta 6 102b 87-121 13 Mackinac 6 890 78-100 14 Emmet 6 98bc 87-109 22 Missaukee 6 lOlbc 91-109 23 Gladwin 6 92c 72-114 24 Arenac 6 930 78-107 1976 Re-collections: 22 Missaukee II 4 95bc 91- 98 23 Gladwin II 3 9 ho 92-105 2 Arenac II 4 9 ho 82-101 Range-wide collections: State & Provinces 247 Montana 3 lOlbc 83-116 259 Quebec 3 105b 103-109 285 Alberta 3 106b 94-120 Betula verrucosa Finland Seed orchards 3 143a 143-145 Species hybrids 1 134a ~— Betula gpp. Mongolian birch l 94bc — van's Seed Orchard 1 lOlbc — Overall mean 55.20m ‘l/ Expressed as percent of population means. 2/ Any two valuse with the same letter are not sign- ificantly different (P 0.05) by LSD. 85 The average 4-month height growth of the 48 families was 53.5cm. Individual families, not separately identified in Table 3-6, varied from a low of 400m (72% of the overall mean) to a high of 67cm (121% of the overall mean). Thus, the Baraga county family (19) and the Delta county family (56) grew almost 68 percent as much as the Gladwin county family (197), which was poorest. Relationships found in Experiment I are similar to those found in this re-analysis study. The differences between families were highly sign- ificant (Table 3-7). Table 3-7. Analysis of variance of juvenile height growth;/ of 48 paper birch families 4-months from seed. Source of variation d.f. M.S. .F-value Between blocks 3 313.60 5.55** Between families 47 173.95 3.08** Between populations 7 251.08 1.56 Families within pOpulations 40 160.45 2.84** Between peninsulas l 84.00 0.48 Error 141 56.48 - 1/ The analysis is based on plot means, measurements made on four trees per plot. ** Significant at the l-percent level. Differences between populations (counties) after combining the six individual family means into a population mean were small, and statistically non-significant (Table 3-7). Thus, the Baraga county population in the UP, at only 86 107 percent of the mean grew only 20 percent as much as the Mackinac county population at 89 percent of the overall mean. Again the relationships found here support the results found in Experiment I, in that the variation between two populations is small compared to the variation between families. When family means for the upper and lower Peninsulas of Michigan were combined, the difference between them was only 2.40m or 4.6 percent. Of interest is the amount of genetic variation that exists within a population. No population had less within- population (family) variation than the maximum difference (20%) of the two extreme populations of Baraga and Mackinac counties. Thus, all pOpulations had a greater amount of within-population variation, up to 58 percent for population number 23 in Gladwin county (Table 3-6). The Baraga county population has a 49 percent height difference between the six individual families. In addition, families 26 and 27 differed by 40 percent, with the parent trees in that stand within 50 yards of each other. Analysis of variance shows that families within p0p- ulations are highly significant and thus accounts for 87.3 percent of the total amount of genetic variation, whereas, differences between populations and peninsulas are non- significant (Table 3-7). The results observed for all 74 seedlots are also presented in Table 3-6. The average 4-month height growth 87 of the 74 seedlots was 55.20m. Analysis of variance revealed highly significant differences between seedlots, between populations, and between seedlots within the populations (Table 3-8). Table 3-8. Anal sis of variance of guvenile height growth of 7 seedlots of birch -months from seed. 1/ Source of variation d.f. M.S. F-value Between blocks 3 276.88 4.36** Between seedlots 73 260.89 4.12** Between populations 17 652.48 4.59** Seedlots within pOpulations 56 142.01 2.24** Error 219 63-37 ' - l/ The analysis is based on plot means, measurments made on four trees per plot. ** Significant at the l-percent level. Seedlots within a population accounted for 41.4 percent of the total genetic variation and populations accounted for 58.6 percent. This is due primarly to the fact that this analysis included all 74 seedlots, including the Finnish sources, which grew significantly faster than all the paper birch seedlots. When only the 48 single families from.Miohigan are analyzed separately, populations account for only 12.7 percent of the total genetic variation. Table 3-6 reveals the large height superiority of the Finnish sources of birch. The three B. verrucosa seed- lots averaged 79.0cm or 143 percent of the overall mean of the study. Thus, under supplemental light, all three 88 Finnish seedlots grew almost twice as much as the source from Gladwin county (197), indicating that B. verrucosa may be more responsive to supplemental light than B. papyxifera. Krizek (1972) reported that B. papypifera may be more responsive to light than B. verrucosa, opposite the tendency found in this study. It is entirely possible though that the seedlots of European white birch, obtained from a seed orchard, have been selected for growth rate. The other two species of birch, Mongolian birch and Van's Pines seed orchard stock remained near the overall mean. Included in this analysis are the three widely scat- tered paper birch population, of which three individual seedlots (families) were grown. These populations also revealed wide variations within the pOpulations, with the Alberta seedlots 1133 and 1136 differing by 27 percent. The Quebec population on the other hand had a maximum family difference of only 5 percent, and could be due to the use of only three seedlots per pOpulation (Table 3-6). In order to test the amount of variation in half-sib progenies from year-to-year, eleven families originally collected in 1975 were re-collected again in the fall of 1976. Variations in juvenile growth between the two successive years is essentially random. Overall mean height within these populations ranged up to 19 percent between families 207A and 208A in Arenac county. 89 Between the two successive years, four 1975 families increased in height, one remained unchanged, and six decreased in height. The largest difference between the-two years was for the Gladwin county families (202 and 202A), which grew an average of 9.00m (17 percent) faster in 1975 than the 1976 re-collection, but this difference was not statistically significant. Correlation analysis of the 4-month seedling height with the 1975 and 1976 collections resulted in a non-significant correlation coefficient (r) of 0.41. Correlation analyses of all paper birch families revealed that height growth under the conditions of cont- inuous light and controlled temperatures is apparently not correlated with latitude, longitude, and length of the growing season (Table 3-9). Table 3—9. Correlation analysis between juvenile height growth of paper birch progeny and environmental variables at the seed origin. Height growth rl/ r-Z/ Environmental variable ‘.—_— *- Latitude 0.44 0.60 Longitude 0-59 0.37 Length of growing season -0.04 -0.05 ;/ Based on 8 populations, including 48 Michigan families. 2/ Based on 14 populations, including all 68 paper birch families. * Significant at the 5-percent level. 9O Weak correlation of height with latitude exists when the three populations of Montana, Alberta, and Quebec are added to the analysis. The unusual aspect of this is that the correlation coefficient is positive, meaning that as latitude is increased from south to north, height is inc- reased. This is Opposite of most geographical variation patterns found in many species, including yellow birch (Clausen, 1968b), and may be due to the effect of cont- inuous light. Age-age correlations were performed to determine how early genetic superiority was established. Measurements were taken at ages of 2%—, 3-, 3%, and 4-months from seed averaging 7.2, 20.2, 39.5, and 55.2cm in height, respectively. As early as Zt-months from seed at an average height of only 7.2cm, a highly significant correlation coefficient (r) of 0.84 was obtained. At 3-months of age r=0.88. Under the conditions of continuous optimum growth, remarkable growth rates are obtained. After 5-months from seed some individual seedling progeny of paper birch and EurOpean white birch, growing continuously, attained over 1500m heights. Although some of the variation between families seem to be due to differences in growth rates, these differences do not follow any consistent geographical pattern. Whether the pattern of variation, if any, in height growth in paper birch is actually random should become clearer with long- term observations and future plantation measurements. 91 Summary and Conciusions The individual parent-progeny data obtained in this study allows an assessment of the amount of genetic var- iation that exists between- and within-populations of paper birch. This information is valuable because it gives the tree breeder an idea as how to implement a selection program to improve paper birch. The results of this study have shown that within-pop- ulation variation in height of 4- and 5-month-old seedlings of paper birch is greater than differences among widely separated pOpulations. Some populations exhibited more family variation than others. Height differences between Michigan populations accounted for approximately 13 percent of the total genetic variation, whereas within-pOpulation (family) variation accounted for 87 percent. Thus, the maximum variation in height growth between any two populations is small (20%) compared to the maximum variation observed between twO families (68%) . Comparative analysis between paper birch and EurOpean white birch revealed the height superiority of the Finnish sources of birch. In the re-analysis section, the four sources, randomly chosen, averaged 143 percent of the over- all mean, growing significantly faster than any families of paper birch, and growing twice as much as the poorest Michigan family (197). Continued observations of European 92 white birch in our established plantations may result in recommendations for its use for reforestation purposes in Michigan. Considerable morphological variability in leaf size, shape, number, and stem pubescence was observed both between- and within-populations of paper birch, which in part reflects its wide geographic range overlapping with many other species of birch. No attempt was made to measure the variation here, but the following chapter will study the characters in more detail. This study provides no evidence of clinal variation in height growth of paper birch. Under greenhouse conditions, no height superiority is established for sources of southern origin, in contrast to the general pattern for many conifer and hardwood tree species of North America (wright, 1976). Certain seedlots grew much better than others, but the variation appears to be essentially random, with much local variation. Although the variation between seedlots seems to be due mainly to differences in growth rates, these differences seem to follow no geographic pattern. Correlation analyses showed that height growth of 4- to 5—month-old seedlings grown under greenhouse conditions apparently are not correlated with latitude, longitude, or length of the growing season. Whether any pattern of var- iation in height growth is actually random should become clearer after long-term observations in test plantations. 93 From a practical standpoint, maximum genetic improve- ment for height growth would be most efficient by practicing family selection. Results from the between- and within- population analysis show there are large amounts of var- iation in height within populations to make individual tree selection worthwile, and demonstrates the importance of progeny testing all paper birches to be used in an improvement program. Thus, much potential growth may be lost if a slow-growing seed source is used for reforestation purposes. CHAPTER IV DEVELOPMENTAL AND MORPHOLOGICAL CHARACTERS OF SEEDLINGS Introduction During the study of genetic variation in growth rates of seedlings of paper birch (Betula papyrifera) reported in the previous chapter, a wide range of morphological var- iability was observed in the seedlings. Since variation in a species is a key prerequisite to its improvement, we must know how much genetic variation is present, and whether or not this variation fits any geographic pattern. As reviewed earlier, several variation studies have been done in yellow birch (Betula alleghaniensis). Morph- ological variation has been studied in characters such as leaf size and shape, catkin, and fruit characters (Clausen, 1968a: Dugle, 1966; Dancik and Barnes, 1972). Some morph- ological studies have been done on paper birch by Brittain and Grant (1966, 1967) reporting their observations of Canadian white-barked species of birch. In their work they examined ten morphological characters of leaves, catkins, bracts, seed, and guard cell length. Their major objective was to define a "typical" white birch. 94 95 In the past, Fernald (1945) recognized a number of varieties. In trying to classify these variants they found they did not invariably correspond to the definition of any named variety. Thus, Brittain and Grant established an arboretum, where a comparison of juvenile characters, growth, and dev— elopment could be carried out. Results of Brittain and Grant's work indicate that considerable variability exist in paper birch due to its wide geographical distribution and occupancy of so many diverse sites. Geographical variants were noted but these were not sufficiently distinct to warrant taxonomic delineation. A number of hybrids were detected between B. papyxifera x B. populifolia. x B. cordifolia, and x B. fontinalis. The purpose of this study is to elucidate the patterns of variation in leaf morphology and develOpmental characters between individual populations of paper birch and seedlots of European white birch. The reson for seeking knowledge of the patterns and underlying causes of this variation, is to relate morphological and developmental characters of the seedlings to the geographic location of the seed source. Thus, some of these differences may reflect a physiological adaptation to a specific geographical area or location. MateriaBs and Methods Materials used for this morphological and developmental study were randomly selected from the previous collections 96 of paper birch and EurOpean white birch, and were described in the previous chapters. A total of 16 seedlots of birch were studied. These included 12 populations randomly selected from 12 widely separated geographic areas over the range of paper birch. Each population consisted of a bulked sample of 6-10 trees per population (stand). In addition, 2 seedlots of B, verrucosa seed orchard material of Finnish origin,and 2 birch hybrids B. papyxifera x‘B. verrucosa (#14) and B. verrucosa x.B. verpucosa (#13) were included in the study. Exact locations of the paper birch populations are given in Figure 4—1. The material was grown from seed. On June 25, 1976 all 16 seedlots of birch were sown in our greenhouse at M.S.U.'s Tree Research Center. The seeds were sown in polycoated (Zieset) paper plant bands 7.50m x 7.5cm x 28cm deep con- taining a soil mixture of 3-parts peat-vermiculite (Redi- Earth) and 1-part coarse perlite. The plant bands were packed in milk cases, each case containing 16 bands. Greenhouse temperatures were maintained in the 210- 27°C range, both day and night throughout the 3%—month growth period. Details of the cultural treatments and handling are identical to those discussed in Chapter III. Seedlings were grown under continuous supplemental light provided by high-output (HO), cool white, 8 foot fluorescent light which produced approximately 120 ft- 97 Figure 4—1. Location and pOpulation number of the 12 paper birch populations used for this study. 98 AP: 3.3.— nausea-.3095 99 candles at the soil surface. The experimental design consisted of a 4x4 square lat— tice, with 5 replications, 4 blocks, and 5 arrangements. Each pOpulation is therefore represented by 4 trees per replication with 5 replications, giving a total of 20 trees per population that measurements were made on. Eighteen characters, encompassing growth parameters, stems, and leaves were measured for each individual seed- ling grown in the greenhouse study. Eight additional var- iables (8,9,10,16,17,24,25,26) were generated as ratios of other measured characters, giving a total of 26 characters as listed below: 2. O\OCD\] O\U\ PM») 11. 12. l3. 14. 15. l6. 17. 18. 19. 20. 21. 22. Growth Parameters Height growth Diameter Stem dry weigh Root 1. l0 Leaf n n Shoot " " Total N I! Root/Shoot ratio LWR (leaf weight ratio) Leaf weight/Total plant D.W. RWR (root weight ratio) Root D.W./Total plant D.W. Stem characters Number of lateral branches Number of internodes Leaf Characters Number of leaves on main shoot Total number of leaves Total leaf area Average total leaf area (Total leaf area/Total number of leaves) LAR (leaf area ratio) = Total leaf area/Total plant D.W. Average area of a single leaf Total leaf length Blade length Petiole length Point of maximum leaf width (W) 100 23. Distance from leaf apex to point of max. leaf width (A). 24. Attenuation factor = A/M ratio 25. Blade length/blade width 26. Number of serrations/average area of a single leaf On October 6, 1976, l3-weeks after sowing, each containerized seedling was measured for height, diameter, number of lateral branches, internodes, and leaves. The' plant bands and growth medium was then carefully washed free of the roots. The stems and roots were collected separately, and packed for oven-dry weight determinations. The leaves of each seedling were measured for leaf area, before oven- drying, with a Lambda Portable Leaf Area Meter Model LI-3000. Individual measurements of leaf characters was performed on a single leaf of each seedling, selected on the basis of the leaf with the largest area per seedling. Actual linear leaf measurements were obtained according to Figure 4-2. Analysis of variance was used to determine the sign- ificance of the 26 characters observed between the 16 seed- lots of birch. Characters were analyzed on Michigan State University's CDC 6500 digital computer, using a square lattice program. 101 Figure 4-2. Diagrammatic sketch of a leaf showing the linear measurement numbers. 103 Results and Discussion Results under the conditions of continuous growth revealed wide variations between the populations of paper birch distributed over its entire natural range. They indicate that few discernible geographic patterns exist for any of the 26 characters measured. It appears that paper birch populations vary essentially in a random manner over the natural range of the species, and follow no set pattern. Roth (1970), working on river birch (B. pxgxx), was also unable to detect any significant clinal variation in leaf width, leaf length, leaf area, or petiole length on either latitude or longitude. Data on all measured develOpmental and morphological characters are presented in Table 4-1. Due to some cultural difficulties encountered in the initial watering of the soil mix, the ten growth parameters showed wide variability but no significant differences. All dry weights measured are directly linked to growth rate of the plant. Thus, any patterns that appear related to height will not be considered further. Of more practical importance are the morphological and developmental characters which are less dependent on growth rate. Statistically significant differences were found for many stem and leaf characters (Table 4-1). Extremely wide variability exists between widely separated pOpulations of paper birch, between paper birch 104 Table 4-1. DevelOpmental and morphological characters measured on 16 seedlots of paper and European white birch. Each value is a mean of 20 seed- lings. _ No. Ave. 188d Origin laggral ifiter— leaves T088? 18:81 total No. branches nodes on main leaves area leaf shoot area - _ -0m2 _ _ _ 22221121224 248 Alaska 1.6 15.9 10.9 13.3 319 21.4 247 Montana 4.7 20.7 15.0 20.8 550 20.8 261 Alberta 5.4 17.7 12.5 19.2 314 19.2 282 Saskatchewan 7.1 19.8 15.1 25.1 314 12.1 267 W. Ontario 2.3 17.7 13.0 14.2 326 19.7 251 N. Minnesota 1.6 15.9 11.1 12.7 262 18.0 252 S. Minnesota 2.0 14.4 9.7 13.1 293 16.5 15 Michigan(UP) 2.4 15.1 [11.2 14.0 399 26.1 16 Michigan(LP) 3.5 16.6 11.8 13.5 237 18.9 272 E. Ontario 2.4 15.1 11.1 13.8 388 23.8 257 Quebec 2.2 14.6 10.2 12.3 234 19.9 291 Mass. 2.0 16.1 9.9 11.3 261 23.0 mean (2) 3.1 16.6 11.8 15.3 325 19.9 European white bixgg 1459 Finland 6.2 20.3 14.4 19.1 347 16.7 1482 Finland 2.8 21.7 15.4 18.0 370 18.3 8001 Finland 1.1 23.1 16.9 19.1 317 15.5 9530 Finland 4.8 20.0 14.5 17.4 348 19.0 mean (2) 3.7 21.3 15.3 18.4 345 17.4 Overall mganl(xg 3.3 17.8 ** 12.7 *416'1 * 330 19.2 ** F-value :oi;2:Z7 2.57 4.31 3.49 2.33 1.28 2.54 LSD (0.05) 3.2 3.8 3.4 7.1 191 6.6 105 Table 4-1. Continued.... Seed- . - sineie Total Blade Petiole ¥§:. Origin LAR gig: izggth length length -cm2 - ------ cm ------ Paper birch 248 Alaska 284 54.2 12.8 9.8 3.0 24? Montana 219 52.9 12.3 9.6 2.7 261 Alberta 266 36.6 10.4 7.8 2.6 282 Sask. 182 30.0 9.3 7.2 2.1 267 w. Ontario 209 ' 49.4 12.1 9.2 2.8 251 N. Minn. 252 44.9 11.9 9.1 2.8 252 s. Minn. 300 42.2 10.7 8.2 2.6 15 Mich. (UP) 211 60.3 13.6 10.7 2.9 16 Mich. (LP) 250 54.1 12.5 10.0 2.5 272 E. Ontario 266 61.2 14.2 11.0 3.1 257 Quebec 231 47.6 12.2 . 9.4 2.8 291 Mass. 301 50.3 12.9 9.8 3.1 mean (i) 248 48.6 12.1 9.3 2.8 European white birch 1459 Finland 200 44.6 11.6 8.5 3.1 1482 Finland 198 45.3 12.0 9.2 2.8 8001 Finland 167 35.0 9.7 7.4 2.3 9530 Finland 199 43.5 11.5 8.6 2.9 mean (2) 191 42.1 11.2 8.4 2.8 Overall mean (2)234 47.0 11.9 9.1 2.8 F-value :8f:%:zg 2.31* 3.36** 5.88** 5.27** 3.17** LSD (0.05) 78 13.4 1.5 1.3 0.46 106 Table 4-1. Continued.... iggd- Origin 5.2.111? 12258.33. $233.. fgggh 32.1332“) No. _leaf point of max.factor /width ave.leaf w1dth(M) Width. (A) (A/M) area ----- cm - - - - - - cm - Paper birch 248 Alaska 7.9 7.3 0.92 1.24 6.6 247 Montana 7.5 7.5 0.99 1.27 7.8 261 Alberta 6.6 6.0 0.90 1.18 9.0 282 Sask. 6.1 5.4 0.88 1.19 9.1 267 W. Ontario 7.3 7.4 1.00 1.25 7.1 251 N. Minn. 7.1 7.0 0.98 1.28 7.1 252 S. Minn. 7.0 6.2 0.89 1.18 7.3 15 Mich. (UP) 7.8 8.3 1.05 1.37 7.7 16 Mich. (LP) 7.5 7.8 1.04 1.34 6.4 272 E. Ontario 8.1 8.4 1.04 1.36 6.7 257 Quebec 7.2 7.0 0.98 1.31 7.7 291 Mass. 7.4 7.5 1.01 1.35 7.8 mean (2) 7.3 7.2 0.97 1.28 7.5 European white birch 1459 Finland 7.8 6.4 0.82 1.09 7.4 1482 Finland 7.6 6.9 0.89 1.19 8.1 8001 Finland 7.0 5.4 0.77 1.05 9.0 9530 Finland 8.0 6.3 0.78 1.08 8.1 mean (E) 7.6 6.3 0.82 1.10 8.2 Overall mean (E) 7.4 6.9 0.94 1.23 7.7 F—value :8{:%:Z$ 1.82 4.42H 10.64** 12.70** 1.77 LSD (0.05) 1.1 1.2 0.08 0.08 1.8 107 and EurOpean white birch, and between the seedlots of European white birch. As stated previously, the paper birch characters studied varied in an apparently random manner, following no general geographic pattern. Of interest though are a few distinct relationships between some of the pOpulations. The most distinctive, is the close relationship of the two paper birch populations from the Canadian Provinces of Alberta (261) and Saskatchewan (282). These two populations, as a group, show many morphological characters significantly different from the other ten paper birch populations. For example they differ in the number of internodes, number of leaves per main stem, total number of leaves per seedling, and the highest number of serrations per average area of a single leaf (Table 4-1). ‘ The Alberta (261) and Saskatchewan (282) populations, particularly the latter, exhibit the lowest average area of a single leaf of all the seedlots (Table 4-l). Also, they develOped the shortest total leaf length, blade length, petiole length, and the narrowest leaf width (M) and the shortest distance from the leaf apex to the point of maximum leaf width (A). Subsequently the leaf attenuation factor, referred to as the A/M ratio (Brayshaw, 1966), was smallest due to the small length of the leaf (A) in com- parison to (M). Thus, in comparison with the other paper birch pOpulations, these populations had a wider leaf width than leaf length, measured from the apex to the point of 108 maximum leaf width (A). These two populations from the northern plains closely resembled g. verrucosa in many developmental respects, such as lateral branching, number of internodes, number of leaves per main stem, and total number of leaves per seedling. In addition, at this early age, they appear similar in stem form, leaf color, and pubescence. 0f unique consequence is the complete absence of stem pubescence. The seedlings from these two populations had no stem pubescence, but rather‘ produced a glandular exudate, Which appeared warty in nature. Brittain and Grant (1975) reported similar observations on young paper birch seedlings from the Rocky Mountain area of Alberta. It is possible that these seedlings are natural hybrids with the small tree or shrub p. fontinalis, which overlapps the range of paper birch in that area. Dugle (1966) has reported that these two species due in fact sometimes hybridize, which may account for the seedlings warty con- dition, commonly found on the twigs of g. fontinalis. The only other evidence of geographic variability in the paper birch pOpulations is shown by the three p0p- ulations from the central Lake States, including pOpulations from Michigan (15 and 16) and eastern Ontario (272). These three are as closely related as any of the populations, especially the UP (15) and eastern Ontario (272) populations. In terms of stem characters no significant differences exist between them and the other paper birch populations. In terms of morphological characters of the leaves however, 109 the three central Lake States populations stand out. These three pOpulations had extremely large leaves, averaging 58.5 square centimeters per leaf, with the eastern Ontario (272) population averaging 61.2 cm2 or over twice the size of the Saskatchewan population (282) which was 30.0 cm2 (Table 4-1). Due to the large leaf size, measurements of blade and petiole length were also larger. To account for this effect, a leaf attenuation factor (A/M) was determined. It is believed that leaf shape is more strongly controlled genetically than leaf size (Crane and Lewis, 1949). Even though the central Lake States populations had wider leaves (M) and a longer distance from the apex to the point of maximum leaf width (A) than other populations, the A/M ratio was consistently large. This indicates that the leaves were proportionately longer than wide. Thus, the Lake States populations had very large leaves in terms of area and they were long and narrow in shape. Research on yellow birch has revealed results similar to those reported here on paper birch. Variation studies of branch and leaf characters showed significant provenance differences, but the variation was small and appeared to be random in most cases. An exception was leaf shape, which appeared to show clinal variation with southern provenances having narrower leaves than northern ones (Clausen, 1973). Dancik (1971) also found Appalachian Mountain trees had narrower leaves than those from the western Great Lakes region. llO The results of my work on paper birch do not substant- iate the trend found in yellow birch. The most southerly source (252) from southern.Minnesota had next to the lowest A/M ratio and the lowest ratio of blade length to width of all twelve paper birch populations (Table 4-1). Two of the more northerly populations from the northern plains of Canada (261 and 282) had the lowest A/M and leaf length/width ratios, supporting the observations on yellow birch. The group with the narrowest leaves were the three populations from the central Lake States. This initial study was designed to provide information on broad geographical variation patterns using a relatively small number of seed sources so that only major types of variation could be detected. An intensive study of a smaller geographic area including many more populations and individual families within these populations is needed to test regional and local variation in a number of characters. This would bring to light any geographic patterns that may actually be present. Included in this study were four seedlots of Finnish origin, two of which (1459 and 1482) are EurOpean.white birch (g. verpucosa) from southern Finland. The other two seedlots are controlled crosses, one is a seedlot obtained from the hybridization of two g. verrucosa provenances (8001), and the second, is a hybrid of g. papygifera x g. verrucosa (9530). 111 The lower portion of Table 4-1 includes the four Finnish seedlots and their respective means for each character. Some of the important differences within the Finnish seed- lots and between them and the.paper birch populations will be discussed. As may be expected, wide variations exist within the four seedlots due to the inclusion of the hybrids. One interesting observation involves the amount of variation in number of lateral branches between the two European white birch provenances, 1459 and 1482, which are within 35 miles of each other in southern Finland. Seedlot 1459 produced over twice the number of lateral branches per seedling as seedlot 1482, significant at the 5-percent level (Table 4-1). The two seedlots were similar in all the other chacters except the leaf shape ratios, A/M and blade/width. The provenance hybrid (8001) showed extreme develop- mental and morphological variability. Compared with all other seedlots this one exhibited very little branching (l.l branches per seedling), the highest number of inter- nodes, largest number of leaves per main stem, lowest LAR, shortest length from leaf apex to the point of maximum width (A), and the lowest A/M and blade length/width ratios (Table 4—1). The most distinguishing characters of the Finnish seed- lots as a group compared with paper birch populations as a group, are the larger number of internodes, larger number of leaves per main stem, lower LAR, and the smaller leaf 112 shape ratios (Table 4-1). In many characters, overlapping frequently occurred, especially with the Alberta (261) and Saskatchewan (282) pOpulation. The Finnish source 1482 was comparable in many characters to paper birch,~including number of lateral branches per seedling, average total leaf area, average area of a single leaf, and all the morphological leaf measurements (Table 4-1). In comparison the paper birch pOpulations and the hybrid g. papygifera x g. verrucosa, very few significant differences exist. This hybrid seems intermediate in most characters, including average total leaf area, average area of a single leaf, total leaf length, blade length and width, petiole length, and number of serrations per single leaf based on area. Large differences do exist between the hybrid and most of the paper birch populations in number of inter- nodes and the leaf shape ratios. The hybrid leaves are noticeably wider and shorter than paper birch leaves, very characteristic of g. verrucosa. Thus, the leaf shape is the only character that is significantly different from all the paper birch seedlots. A comparison of the hybrid birch with the two g. 32;- rucosa seedlots reveals no significant differences in any character measured. The hybrid appears to resemble European white birch more than paper birch. Johnsson (1949) studied morphological characters of birch hybrids and found that the hybrid g. verrucosa x g. papyrifera is intermediate in leaf size. It is possible that my results do not show the 113 intermediate relationship because I used only two 2. 1g;- rucosa seedlots. The use of more seedlots of the species and hybrid birch might show the hybrid to be intermediate. Casual observations were made on stem pubescence but it was not measured. It appears that stem pubescence varies also in an apparently random manner. The stem 0f paper birch is usually densely pubescent beginning in the early seedling stage. Some of the paper birch pOpulations differ from others, but four populations had seedlings which varied in pubescence within the population. The Alaskan source (248) differed from other populations in developing only slight pubescence. In addition, variation also existed within this population, with about 25 percent of the seed- lings deve10ping a glandular condition on the stem and complete absence of pubescence. This same trend with respect to stem pubescence occurred on populations from western Ontario (267), northern Minnesota (251), and eastern Ontario (272). The reason for this variation is probably because the species occupies such diverse sites and has a wide geographic range. Its range is sympatric with g. neoalaskana, g. cordifolia, and g. populifolia all of which exhibit the glandular exudate at an early age. This probably allows numerous varieties of paper birch to develOp through hybridization. It is interesting to note that within a population, consisting of about ten trees bulked together, variation in stem pubescence also exists between individual parent trees in a 114 stand differ in stem pubescence. Further studies of other morphological and anatomical characters of individual trees in a stand are needed to clearly illustrate the exact nature of inherited variation. Two populations that show extreme differences in stem pubescence in all its progeny are the Alberta (261) and Saskatchewan (282) pOpulations was entirely lacking stem pubescence, and produced a glandular exudate. The cause of this variation can possibly be ascribed to hybridization with g. fontinalis. Dense pubescence on the other hand occurred consistently in three populations of paper birch, including both Michigan pOpulations (l5 and 16) and the southern Minnesota population (252). Only slight variations in pubescence were observed within these pOpulations. Pubescence was also variable on the Finnish seedlots of birch. EurOpean white birch develOped a moderate pubescence whereas the birch hybrid (8001) showed a wide range of variability from moderately pubescent to a slight glandular state. Suprisingly, the g. verrucosa x g. papyrifera hybrid (9530) had moderate pubescence consist- ently on most of the seedlings. Summary andConclusions Variation in 26 develOpmental and morphological characters was essentially random with respect to geograph- ical gradients over the natural range of paper birch. 115 Significant differences between seed sources were observed for many stem and leaf characters in seedlings of paper birch, European white birch, and the hybrids between them. Extreme variation was found in two populations occurring in the northern plains of Canada, especially in stem pubescence, number of lateral branches, average area of a single leaf, and leaf shape ratios. Paper birch populations from the central Lake States develOped very large leaves in terms of area, and the leaves were long and narrow in shape. Two European white birch seedlots overlapped in many foliar characters with paper birch, differing significantly only in producing larger number of internodes, larger numbers of leaves, and lower leaf shape ratios. The hybrid g. papygifera x g. verrucosa was intermediate in many morph- ological characters except for leaf shape which closely resembled European white birch. Stem pubescence was also variable and ranged from a glandular form to densely pubescent. No geographical patterns were observed except for the distinctness of the two populations in the northern plains. It is entirely possible that much of the variation in pubescence may be due to natural hybridization of paper birch and other birches with sympatric ranges. No significant clinal variation associated with latitude or longitude was found in any of the developmental and morphological characters measured. 116 There was an absence of large morphological and develOpmental differences between the Alaskan population (248) located the farthest north and west, and the other pOpulations further east and south. For example, the Quebec (257) and Alaskan populations are separated by over 2,000 miles but did not differ significantly in any character measured. In addition, the most southern source, located in southern Michigan (16), differed significantly from the Alaskan pOpulation in only the two leaf shape characters. Although this rangewide study on widely separated populations provided information on gross geographic patterns in morphology and development, it also brought out the need for further research on local variation patterns within an individual population. Variations in stem pubescence and juvenile height growth, as demonstrated in Chapter III illustrate the need for intensive studies over small geo- graphic areas. These studies should include many more populations with data on progeny performance of families selected within these populations. Information on local variation patterns for a number of characters within a population could reveal the relative amount of variation due to gene action and a complex of environmental variables. Thus further investigations of paper birch are needed in order to better understand the causes of the intraspecific variability found in this research. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS Paper birch seed and seedlings revealed a large amount of geographic variation which was not related to any en- vironmental gradients. Seed germination The germination capacity and germinative energy of paper birch seed was significantly affected by temperature of the test conditions and geographic origin. Seed from the Upper Peninsula (UP) of Michigan had a wider range of temperatures Optimum for germination, and increased germinative energy, in comparison with seed from the Lower Peninsula (LP) of Michigan. In addition, the UP seed revealed much plasticity in its ability to germ- inate at very low (10°C) and high (43°C) temperatures. Seed from the UP initiated seed germination earlier than seed from the LP, especially at lower temperatures. Germination capacity was found to vary significantly both within- and between-pOpulations of paper birch. Within a population, individual families occassionally differed by 70 percent or more in percent germination. The amount of seed borne by the parent tree, or seed crop, showed no correlation with percent germination of the seedlot, and seed cr0p varied both within- and between- ll? 118 pOpulations. One population with the lowest average seed crop rating of the parent trees had one of the highest germination percentages. Variation in seed germinative energy was also related to geographical origin. The combined UP populations reached 63 percent germination in 7 days, whereas the combined pOpulations from the LP took nearly 13 days to obtain the same percent germination. The most distinctive difference in seed germination was noted between the western UP and south-central LP pop- ulations. Seed from pOpulations in the western UP began germination sooner, proceeded more rapidly, and reached optimum levels earlier than populations in the south-central LP. The rapidity of germination or germinative energy was not correlated with size of the parental seed cr0p. Growth Variation in height growth of 4- and 5-month-old green- house-grown seedlings of paper birch, was due mostly (87%) to within-pOpulation differences. In other words variation between half-sib families was greater than differences among widely separated populations. The amount of variation in growth between families differed among the pOpulations sampled, but the maximum variation between any two populat- ions was small (20%) compared to the maximum variation observed between two families (68%). 119 Genetic variation in juvenile height growth of container-grown paper birch seedlings grown under green- house conditions was essentially random, with respect to environmental gradients. Long-term observations in plant- ations should show whether or not height growth is actually random under natural conditions and over longer periods of time. EurOpean white birch from Finland was superior to paper birch in height growth. If this superiority holds true in test plantations it would indicate a possible use of European white birch for reforestation purposes in Michigan. Morphology Considerable morphological variability in leaf size, shape, number, and stem pubescence was observed both between- and within-pOpulations of paper birch. Variation in 26 developmental and morphological characters measured was essentially random over the natural range of paper birch, following no geographic or clinal pattern. Significant amounts of genetic variation were found in a number of characters, including number of lateral branches, stem pubescence, leaf number, leaf area, leaf length, and leaf shape. Variability in the above characters exists both between widely separated populations of paper birch, and between paper and EurOpean white birch. 120 Two distinct ecotypes were found in the range-wide morphological study of paper birch. The populations from the northern plains of Canada (Saskatchewan and Alberta) and from the central Lake States as a group, showed similarities within groups in many of the characters examined, separating them from the other pOpulations. The two European white birch seedlots overlapped with paper birch in many foliar characteristics, but differed significantly in production of larger numbers of internodes, larger number of leaves, and low leaf shape ratios. Recommendations This study of widely separated paper birch populations provided information on broad geographic variation patterns in germination characteristics, juvenile height growth, and morphology and develOpment. It also brought out the need for further research on local variation patterns within an individual pOpulation. Variations in stem pub- escence and juvenile height growth illustrate the need for intensive studies within small geographic areas. These studies should include many more populations with data on progeny performance of families selected within these populations. Further genetic studies of paper birch are required to learn more about intraspecific variability, in order to serve as a basis for future breeding strategy. There is also a need to establish correlations between juvenile and mature physiological and morphological characteristics 121 of paper birch. Information is needed on juvenile height growth of half-sib progenies grown under lights versus natural photo- periods in a greenhouse environment. Such studies would determine the effect of supplemental light on the per- formance of the seedlings, and the residual effects of these treatments on the trees in test plantations. Growth rate, branching, and branch angle are important characters for the improvement of tree quality and increasing the financial return of plantation-grown paper birch. The considerable amount of genetic variation in these characters revealed in this study indicate a potential for their improvement. Observations of European white birch should be cont- inued in our established plantations because they could result in recommendations for its use for reforestation in Michigan. From a practical standpoint, maximum genetic improve- ment for height growth would be accomplished most efficiently by practicing individual tree (family) selection, due to the large amounts of within-population (family) variation present. Results indicate the importance of progeny testing all paper birches to be used in an improvement program, without progeny testing much potential growth improvement may be lost if slow-growing seed sources are used for reforestation purposes. BIBLIOGRAPHY Bjorkbom, J.C. (1971). Production and germination of paper birch seed and its dispersal into a forest opening. USDA Forest Service Res. Paper NE-209, 1 pp. NE For. Exp. Sta., Upper Darby, PA. Bjorkbom, J.C., D.H. Marquis, and F.E. Cunningham. (1965). The variability of paper birch seed production, dispersal, and germination. USDA Forest Service Res. Paper NE-4l, 8pp. NE For. Exp. Sta., Upper Darby, PA. Black, M., and P.F. Wareing. (1954). Phot0periodic control of germination in seeds of birch (Betula pubescens). Nature 174x705—706. Black, M., and P.F. Wareing. (1955). Growth studies in woody species VII. Phot0periodic control of germ— ination in Betula pendula. Physiol. Plant. 8:300-316. Blyth, J.E., and J.T. Hahn. (1974). Pulpwood production in the North Central region by county, 1974. USDA Forest Service Resour. Bull. NC-29, 26pp. NC For. Exp. Sta., St. Paul, MN. Boyer, W.D. (1972). Air temperature, heat sums, and pollen shedding phenology of longleaf pine. Ecology 54 x “20-426 0 Brayshaw, T.C. (1966). What are the blue birches? can. Field-Natllro 803187‘1940 Brittain,W.H., and W.F. Grant. (1966). Observations on Canadian birch (Betula) collections at the Morgan Arboretum. III. B. ifera of British Columbia. gap. Field-Natur. 80:1 7-157. Brittain, W.H., and W.F. Grant. (1967). Observations on Canadian birch (Betula) collections at the Morgan Arboretum. V. B. papypifera and B. cordifolia from eastern Canada. gap. Field-Natur. 81:251-262. Clausen, K.E. (1962). Introgressive hybridization between two Minnesota birches. Silvae Genet. 11:142-150. 122 123 Clausen, K.E. (1968a). Natural variation in catkin and fruit characteristics of yellow birch. Proceedipgs of the Fifteenth NE Forest Tree Improv. Conf., pp2-7. Clausen, K.E. (1968b). Variation in height growth and growth cessation of 55 yellow birch seed sourcces. Eighth Lake States For. Tree Impr. Conf. Proc., ppl-4. Clausen, K.E. (1972). Within—Provenance variation in yellow birch. Twentieth Northeastern For. Tree Im r. Conf. Proc., pp90-98. U. of New Hampshire, 1972. USDA For. Serv., NE For. Exp. Sta., (1973), l77pp. Clausen, K.E. (1973). Genetics of yellow birch. USDA Forest Service Research paper WO-l8, 28pp. Clausen, K.E., and P.W. Garrett. (1969). Progress in birch genetics and tree improvement. In Birch Symposium Proceedings : 86-94. USDA For. Serv., NE For. Exp. Sta., Upper Darby, PA. Conklin, J.G. (1969). Insect enemies of birch. Ip Birch Symposium Proceedings:151-l54. USDA For. Serv. NE For. Exp. Sta., Upper Darby, PA. Crane, M.B., and D. Lewis. (1949). Genetical studies in pears. Heredipy 3:85-97. Dancik, B.P. (1971). Multivariate analysis of variability in leaf morphology of yollow birch (Betula alleghan- iensis Britton) in the western Great Lakes region. PhD. Thesis, Univ. of Michigan, Ann Arbor, MI., 65pp. Dancik, B.P., and B.V. Barnes. (1972). Natural variation and hybridization of yellow birch and bog birch in southeastern Michigan. Silvae Genet. 21:1-9. Dugle, J.R. (1966). A taxonomix study of western Can- adian species in the Genus Betula. Can. 1. Bot. “'48 929-1007 0 Duncan, D.B. (1955). Multiple range and multiple F tests. Biometrics 1131-42. Fernald, M.L. (1945). Some North American Copylaceae (Betulaceae . In: Notes on Betula in eastern North America. Rhodora 47:303-329. Fowells, H.A. (1965). Silvics of U.S. Forest Trees. U.S.D.A. Forest Service, Agr. Handbook #271, 762pp. 124 Fraser, J.W. (1971). Cardinal temperatures for germination of six provenances of white spruce seed. Department of Fisheries and Forestry, Canadian Forestry Service, Publ. #1290, 10pp. Govorukha, G.I., and S.A. Mamaev. (1971). Effect of temp- erature regimes on the viability and germinative energy of seeds of Betula verrucosa, and B. pubescens of different geographic origins. Ekologiya 3: 7-52. Grant, W.F., and B.K. Thompson. (1975). Observations on Canadian birches, Betula cordifolia, B. neoalaskana, B. p0pulifolia, B. papyrifera, and B. x caerulea. Can. J. Bot. 53:1478:1490. Halliday, W.E.D., and A.W.A. Brown. (1943). The distrib— ution of some important forest trees in Canada. Ecology 24:353-373. Harlow, W.M., and E.S. Harrar. (1969). Textbook of Dendrology. 5th ed. McGraw-Hill, Inc., New York. Hepting, G.H. (1971). Diseases of forest and shade trees of the U.S. USDA Forest Service Agr. Hdbk. #386, 658pp. Horn, A.G. .(1966). Lake States pulpwood production hampered by adverse weather and labor shortage, 1965. U.S. Forest Service Res. Note NC-3, 2pp. North Central Forest Exp. Sta., St. Paul, MN. Hutnik, R.J., and F.E. Cunningham. (1961). Silvical characteristics of paper birch (Betula papypifera). U.S. Forest Service Station Paper ¥14 , 2 pp. NE For. Exp. Sta., Upper Darby, PA. Hyvarinen, M.J. (1968). Paper birch, its characteristics, properties, and uses: a review of recent literature. U.S. Forest Service, Res. Paper NC-22, 12pp. NC For. Exp. Sta., St. Paul, MN. Johnsson, H. (1945). Interspecific hybridization within the Genus Betula. Hereditas XXXI. Johnsson, H. (1949). Studies on birch species hybrids I. Betula verrucosa x B. japonica, B. verrucosa x,§.,pap- yplfera, and B. pubesceng X.§- papypifera. Hereditas XXXV. Krizek, D.T. (1972). Accelerated growth of birch in control- led environments. Proc. Int. Plant PrOp. Soc. Ann. Mtg. PP390~394- . . Maini, J.S., and E.S.P. Wang. (1967). TWln seedlings and abnormal germination in yellow birch, Betula allegh- aniensis Britten. Can. Field-Natur. 81:128-134. 125 Marquis, D.A. (1969). Silvical requirements for natural birch regeneration. Birch Symposium_§roceedin s, : 40-49. USDA Forest Service, NE For. Exp. Sta., Upper Darby, PA. Marquis, D.A., D.S. Solomon, and J.C. Bjorkbom. (1969). A silvicultural guide for paper birch in the North- east. USDA Forest Service Res. Paper NE-l30, 47pp. NE For. Exp. Sta., Upper Darby, PA. Mergen, F. (1963). Ecotypic variation in Pinus strobus L. Ecology 44:716-727. McDermott, R.E. (1953). Light as a factor in the germ- ination of some bottomland hardwood seeds. i. For. 513203-204. McNaughton, S.J. (1966). Ecotypic variation in the Typha community-type. Ecol. Monog. 36:297-325. Phillips, W.D. (1922). Effect of climatic conditions on the blooming and ripening dates of fruit trees. Cornell Univ. Agric. Exp. Sta. Memo 59:1379-1416. Quigley, K.L., and H.M. Babcock. (1969). Birch timber resources of North America. Birch S osium Pro- ceedin s, :6-14. USDA Forest Serv1ce, NE For. Exp. Sta., Upper Darby, PA. Roth, P.L. (1970). Phenotypic variation in river birch, Betula nigpa L. Proceedings of the Indiana Academy piScience. 80:225-229. U.S. Forest Service. (1974). Seeds of woody plants in the U.S. USDA Agric. Handb. #450. Vaartaja, 0. (1952). Forest humus quality and light con- ditions as factors influencing damping-off. Phytopathology 42:501-506. Vaartaja, O. (1959). Phot0periodic response in germ- ination of seed of certain trees. Can. J. Bot. 34' 377-388 I Winstead, J.E. (1971). Populational differences in seed germination and stratification requirements of sweetgum. For. Sci. 17:34—36. Wright, J.W. (1976). Introduction to Forest Genetics. Academic Press. New York, San Francisco, London. lllllllHllllll H S R E N N U E T A” MHZ 1 3 umnmum