Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Genetic differences in physiological and morphological characteristics of Sitka alder populations in… Benowicz, Andy 1998

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1998-0356.pdf [ 5.78MB ]
Metadata
JSON: 831-1.0088530.json
JSON-LD: 831-1.0088530-ld.json
RDF/XML (Pretty): 831-1.0088530-rdf.xml
RDF/JSON: 831-1.0088530-rdf.json
Turtle: 831-1.0088530-turtle.txt
N-Triples: 831-1.0088530-rdf-ntriples.txt
Original Record: 831-1.0088530-source.json
Full Text
831-1.0088530-fulltext.txt
Citation
831-1.0088530.ris

Full Text

GENETIC DIFFERENCES IN PHYSIOLOGICAL AND MORPHOLOGICAL CHARACTERISTICS OF SITKA ALDER POPULATIONS IN BRITISH COLUMBIA by A N D Y B E N O W I C Z B.Sc. (Ag.) Hon., The University of British Columbia, 1995 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF M A S T E R OF S C I E N C E in T H E F A C U L T Y OF G R A D U A T E S T U D I E S ( D E P A R T M E N T OF F O R E S T S C I E N C E S ) We accept this thesis as conforming to the required standard U N I V E R S I T Y OF B R I T I S H C O L U M B I A September, 1998 © Andy Benowicz, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study, i further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of F^E^ST SClfflwCcri The University of British Columbia Vancouver, Canada Date Q2 . Q&JU> , IgS* DE-6 (2/88) 11 A B S T R A C T Patterns of genetic variations in adaptive and quantitative attributes of Sitka alder, Alnus sinuata Rydb., were examined at the population level. The following traits were studied: germination, frost hardiness, bud break, characteristics related to gas exchange (photosynthesis, dark respiration, stomatal conductance, transpiration, carboxylation efficiency, stomatal sensitivity to vapor pressure deficit and intrinsic water use efficiency) and characteristics related to biomass allocation (height, stem diameter, shoot dry weight, root dry weight, height growth rate and number and size of major stems per plant). The number of examined populations was 29, all but one from British Columbia. There were large genetic differences among the populations in all measured traits except for spring frost hardiness and stomatal sensitivity to vapor pressure deficit. Population genetic structure (variance within and between populations) was investigated: inter-population variations accounted for 13-31% and 26-61%) for gas exchange and biomass related variables, respectively. The populations differed significantly in fall and winter cold resistance, though the differences were less pronounced in maximum hardiness than were found for the timing of frost hardiness development. On 11 November, 1996 65% of the total variation in frost injury index was explained by the differences between the populations, while on 15 December, 1996 33%) of the total variance was due to the population effect. The observed patterns of large genetic variations in Sitka alder populations represent a challenge for gene conservation efforts in high elevation ecosystems. Based on univariate (multiple linear regression) and mulitvariate methods (canonical correlations and cluster analysis), geographic patterns of variation were found for most traits. Variations related to geography were particularly strong in fall frost hardiness (R 2 = 0.62), shoot dry weight (R = 0.64), growth rate in July (R = 0.70), ratio of root dry weight to total dry Ill weight (R 2 = 0.67) and net photosynthetic rate (R 2 = 0.52). Most of the variation was due to latitude and distance from the coast. Seedlings generated from northern seed sources were in general more frost hardy, smaller and allocated more carbon to roots compared to shoots. Plants from the interior wet belt were frost hardy to the same levels as coastal seedlings suggesting that snow cover plays a role in Sitka alder adaptation to low temperatures. Rate of photosynthesis (A) increased largely with the distance from the coast and, to a lesser degree, with elevation and latitude implying that an inverse relationship exists between A and the length of the growing season. Interior populations had higher A due to higher carboxylation efficiency and stomatal conductance (gs). A s a result of higher gs, plants from more xeric interior regions tend to have lower water use efficiency. Productivity was negatively correlated with A and positively with water use efficiency. Speed of germination did not affect the plant size after one growing season and was independent from the germination completeness. Strong positive correlation was found between timing of frost hardiness development and spring bud break. iv T A B L E OF C O N T E N T S A B S T R A C T i i LIST OF T A B L E S v i LIST OF F I G U R E S v i i LIST OF S Y M B O L S ix A C K N O W L E D G E M E N T " x i 1.0 I N T R O D U C T I O N 1 2.0 L I T E R A T U R E R E V I E W 3 2.1 Species description 3 2.2 Germination 4 2.3 Plant productivity, gas exchange and water relations 6 2.4 Frost hardiness 11 2.5 Repeatability 13 3.0 M A T E R I A L S A N D M E T H O D S 15 3.1 Plant material 15 3.2 Germination 18 3.3 Frost hardiness and bud burst 18 3.4 Gas exchange 20 3.5 Quantitative attributes 22 3.6 (a) Statistical analysis 23 3.6(b) Analysis of variance models and experimental design 26 4.0 R E S U L T S 29 4.1 Germination 29 4.2 Frost hardiness 33 4.3 Gas exchange parameters 40 4.4 Quantitative attributes 51 4.5 Relationship between the data sets 57 5.0 D I S C U S S I O N 61 5.1 Germination 61 5.2(a) Frost hardiness 66 5.2(b) Frost hardiness and bud break. 70 5.3(a) Gas exchange 73 5.3(b) Gas exchange and biomass accumulation 76 5.3(c) Water relations 77 5.4 Quantitative attributes 79 6.0 C O N C L U S I O N S A N D S U M M A R Y 83 V 7.0 F U T U R E D I R E C T I O N S 85 R E F E R E N C E S 87 A P P E N D I X 1 97 A P P E N D I X 2 98 A P P E N D I X 3 99 A P P E N D I X 4 101 A P P E N D I X 5 102 A P P E N D I X 6 105 A P P E N D I X 7 107 A P P E N D I X 8 109 A P P E N D I X 9 I l l A P P E N D I X 10 115 LIST OF T A B L E S Table 1. Population location and symbol. 16 Table 2. One-way analysis o f variance; tp = intraclass correlation coefficient. 26 Table 3. Two-way analysis of variance; tp = intraclass correlation coefficient. 27 Table 4. Analysis of variance table for repeatability. 28 Table 5. Components of variance (a 2) for population FII evaluated at five dates. Data are presented for the test temperature which produced the largest differences among the populations at each date. 34 Table 6. Overall mean, standard deviation (SD), means' range and relative range (RR = range 100/mean), intraclass correlation coefficients (tp) and repeatabilities (Rp) of variables related to gas exchange. Overall mean and standard deviation were calculated using individual observations. Range and relative range were calculated for population means. 41 Table 7. Overall mean, standard deviation (SD), means' range and relative range (RR = range 100/mean), intraclass correlation coefficients (tp) of variables related to morphology. Overall mean and standard variation were calculated using all individual observations. Range and relative range were calculated for population means. 52 LIST OF F I G U R E S Figure 1. Rate of net photosynthesis (A) versus intercellular C 0 2 concentration (Ci). After Farquhar and Sharkey (1982). 8 Figure 2. Map of studied populations of Sitka alder in British Columbia. Part of British Columbia where Sitka alder does not grow is delineated in gray. 17 Figure 3. Cumulative germination of 27 Sitka alder populations during the standard germination tests. 31 Figure 4. Factor loadings plot for germination parameters. 32 Figure 5. Change in mean frost injury index (%) of individual populations evaluated at - 1 8 ° C from 15 October, 1996 to 09 March, 1997. 36 Figure 6. Differences among the populations in frost hardiness development based on dates when each population attained resistance to -18°C. Frost resistance was measured as population mean LT50. 37 Figure 7. Mean population LT50 (°C) on 20 January, 1997. 38 Figure 8. Cluster analysis dendrogram of population grouping based on frost hardiness. Population FII means of all tests and temperatures were used as input variables. 39 Figure 9. Mean rate of net photosynthesis (A) of interior and coastal populations versus temperature. Each mean was calculated based on temperature interval of 1 degree. Error bars represent standard error of the mean; n.s. = non-significant difference (P>0.05). 44 Figure 10. Mean C i / C a of interior and coastal populations versus temperature. Each mean was calculated based on temperature interval of 1 degree. Error bars represent standard error of the mean; s.d. = significant difference (P<0.05). 45 Figure 11. Mean rate of net photosynthesis (A) o f interior and coastal populations versus C i /Ca . Each mean was calculated for the same temperature intervals as in Figure 9. Error bars represent confidence intervals of the means (95%). 46 Figure 12. Mean rate of net photosynthesis (A) of interior and coastal populations versus stomatal conductance (gs). Each mean was calculated for the same temperature intervals as in Figure 9. Error bars represent confidence intervals (95%) of the means. Hyperbolic curves (y=ax/(b+x)) were fitted for both data sets. 47 Figure 13. Mean stomatal conductance (gs) of interior and coastal populations. Error bars represent standard error of the mean; n.s. = non-significant difference (P>0.05). Each mean was calculated based on temperature interval of 1 degree. (A) vapour pressure deficit (VPD) at the leaf surface and (B) transpiration (E) of interior and coastal populations versus temperature. 48 Figure 14. Mean stomatal conductance (gs) of interior and coastal populations versus vapour pressure deficit (VPD) . Each mean was calculated based on V P D intervals of 0.1 kPa. Error bars represent standard error of the mean; s.d.=significant difference (P<0.05). 49 Figure 15. Cluster analysis dendrogram of population grouping based on gas exchange data. Population means of all gas exchange related variables were used as input variables. 50 Figure 16. Changes in mean population height from 5 A p r i l to 25 September, 1997. 54 Figure 17. Changes in mean population height growth rates (cm/day) during the growing season. 55 Figure 18. Mean population root weight ratio (RWR) plotted against the latitude of the seed collection sites. Dashed line is based on regression with all populations excluding Californian population. 56 Figure 19. Mean population germination capacity (GC) and the average number of major stems (including the main stem) per plant (STEMS) . 59 Figure 20. Frost hardiness development and bud break. FII = frost injury index (%) evaluated on 11 November at - 25°C. DAYS5o%budbreak=number of days from 15 March for 50% of population plants to start bud break. (A) map of bud break timing in British Columbia, D A Y S = DAYSsc/obudbreak; (B) map of November frost hardiness in British Columbia. Note: to convert minutes of latitude and longitude to degrees divide by 60. 60 ix LIST OF S Y M B O L S A = net instantaneous photosynthetic rate (pmol CO2 m" 2 s"1) B l = mean conductivity of the two blanks after exposure to low temperature B2 = mean conductivity of the two blanks after exposure to 90 °C C I = conductivity following low temperature treatment (mho) C2 = conductivity of kil led tissues (mho) Ca = ambient CO2 concentration (pL L" 1 CO2) C A L = stem caliper measured at the root-shoot transition zone (mm) C i = intercellular CO2 concentration (pL L" 1 CO2) C O A S T = distance from the coast (km) DATELT50 ( -18) = date of population frost resistance to -18°C expressed as LT50 DAYS5o%budbreak = number of days from 15 March for 50% of population plants to start bud break 2 1 E = transpiration rate (mmol H2O m" s") E L E V = elevation FII = frost injury index (%) G C = germination capacity (%) 2 1 gm = carboxylation efficiency (mol CO2 m" s") G R J = absolute growth rate between 09 July and 12 August (cm/day) gs = stomatal conductance (mmol H2O m" 2 s"1) G U = germination uniformity (%) G V = germination value H I = height after the first growing season (cm) H2 = height after the second growing season (cm) L A = leaf area per plant (cm ) L A T = latitude L O N G = longitude It = leaf temperature (°C) LT50 = lethal temperature that kil ls 50% of tissues (°C) P = population P A I = plant architecture index P V = peak value of germination (number of seedlings/day) R = dark respiration rate (pmol CO2 m" s") r = simple correlation coefficient R = coefficient of determination for multiple linear regression r 2 = coefficient of determination for simple linear regression R '50= relative germination rate (days) R C C = relative conductivity of control sample (%) R C t = relative conductivity of sample exposed to treatment temperature (%) R D W = root dry weight (g) R H = relative humidity (%) Rp = repeatability R W R - root weight ratio SD = standard deviation S D W = shoot dry weight (g) S E M = standard error of the mean a 2 = variance STEM = number of major stems per plant T = transformed TDW = total dry weight (g) tp = intraclass correlation coefficient 2 1 1 VPDsens = stomatal sensitivity to water vapor pressure deficit (sm" mol H2O" kPa") W U E = photosynthetic intrinsic water-use efficiency (pmol C 0 2 /mmol H2O) xi A C K N O W L E D G E M E N T I would like to take this opportunity to express my deep gratitude for help given to me by a number of people who made it possible for me to complete this project. I would like to thank my supervisors, Dr. Yousry El-Kassaby and Dr. Robert Guy for sharing their knowledge and experience, for their continuous support in every possible way during the course of my studies and for their encouragement in more difficult moments. I am also grateful to other members of my supervisory committee, Dr. Christopher Chanway and Dr. Cheng Y i n g for the useful input in the early stages of my research as well as time and effort they took to review the completed thesis. Likewise, I would like to thank Dr. Roy Turkington, the non-departmental examiner, for careful examination of my thesis and helpful comments. Dr. Cheng Y i n g kindly provided seeds for the study. The invaluable technical support of the Saanichton Forestry Centre personnel, especially Cathy Cook, John Halusiak, Debbie MacLeod, M e l Turgen and Catharine Sutcliffe is much appreciated. I would also like to thank my fellow graduate students, staff and the Faculty of Forestry members for various forms of assistance during my studies. This project was funded by Forest Renewal British Columbia grant, Genetics of British Columbia Minor Tree Species, awarded to Dr. Yousry El-Kassaby. Special thanks go to my wife, Ania , and my son, Thomas, for always being there for me. 1 1.0 I N T R O D U C T I O N Human activities reduce the diversity of forest ecosystems directly by changing land use patterns or indirectly, for example, by pollution. Dramatic reductions in the number of species or their population sizes are accompanied by less easily observable erosion of within-species genetic variation. Conservation and responsible use of genetic resources is dependent upon the knowledge of the extent and pattern of genetic intraspecific variance. Population testing is one of the methods employed to provide the needed information. While such tests have been established for many commercially important tree species, less attention has been paid to species with no timber value. Importance of these species, often essential components of healthy ecosystems, can be shown even from a narrow economic perspective. Through more or less explicit interactions, the non-crop plants frequently have positive or negative impact on commercial trees. In addition, there can always be potential economic uses, not related to wood production, with Pacific yew (Taxus brevifolia Nutt.) serving as a case in point. Sitka alder (Alnus sinuata Rydb.) is another good example of a species with no wood value playing a role from the ecological and economic point of view. Often regarded as "a weed", it is now produced for the purpose of soil erosion control and slope stabilization. While it often competes with more valuable species during forest regeneration, it can also have beneficial effects as it is capable of improving site productivity by adding organic matter and nitrogen to the soil. A s the shortage of trees available for harvest at lower elevations becomes a more widespread reality, there is an increasing interest in timber harvesting in subalpine areas where Sitka alder grows. This implies that this species w i l l be more frequently subject to management practices and thus its genetic resources w i l l be affected. Maintaining genetic diversity of higher elevation ecosystems may be particularly challenging due to the harsher climate. A t present, no information is available about the genetic variability in Sitka alder 2 populations in British Columbia. A population trial, covering much of the natural range of the species in the province, was established on Vancouver Island with an objective to examine the structure and patterns of genetic variations in Sitka alder adaptive and quantitative attributes. Another objective of the study was to relate the uncovered patterns of variations to the environmental conditions of seed collection sites to study trends in adaptation. A number of adaptive and quantitative attributes as well as the interdependencies among them were investigated. The attributes were related to germination, gas exchange, frost hardiness and biomass allocation. A n additional purpose o f the experiment was to find the best populations from the management viewpoint (mainly slope stabilization and road deactivation) for the south-coastal region of B . C . and for Vancouver Island, i.e. for the regions where utilization of Sitka alder is l ikely to be most intensive. Based on the wide distribution range of the species in British Columbia along a number of environmental gradients, it was expected that large genetic differentiation at the population level would be found as was the case for many other species (Morgenstern, 1996). In addition, the differences among the populations may be particularly large since Sitka alder may sometimes be subject to the random genetic drift that increases the variance between the populations. This follows from the species' pioneer nature (likelihood of the founder effect), possible site isolation due to the growth in high elevation habitats, and population dynamics characterized by rapid expansion after disturbance (e.g. windthrow, avalanche, fire, logging) followed by reduction in the population size due to interspecific competition. Change in gene frequencies by chance is more probable in small populations and during the establishment of new populations. It can effectively reduce within population variance, thus increasing variance between populations. 3 2.0 L I T E R A T U R E R E V I E W 2.1 Species description Sitka alder is a fast-growing pioneer plant capable of living in disturbed as well as in stable environments. It is a small, deciduous tree or a tall bush with shallow roots (Farrar, 1995) occurring throughout British Columbia except for the northeast corner of the province. It grows at middle to higher forested elevations in cool and moist climates and in every biogeoclimatic zone of British Columbia excluding the Ponderosa Pine and Bunchgrass zones (Krajina et al., 1982). Sitka alder is reported to be a very frost tolerant plant (Haeussler et al., 1990) also capable of withstanding heavy snow and avalanches due to flexible branches (Lyons, 1952). Unable to efficiently control moisture loss from the foliage, it prefers moist sites and is not well adapted to drought (Cline and Campbell, 1976). A s far as light conditions are concerned, Sitka alder is regarded as a moderately shade tolerant species (Krajina et al., 1982) growing well on exposed sites but poorly under dense cover. Sitka alder often forms a symbiosis with Frankia (Actinomycetes), a soil-living bacterium that can utilize atmospheric nitrogen. Rates of nitrogen fixation of 20 to 150 kg/ha per year are similar to the quantity of nitrogen fixed by red alder (Alnus rubra Bong.) as reported by Binkley (1982). This ability allows Sitka alder to colonize sites poor in nitrogen and organic matter (e.g. exposed mineral soil). It is often found on clearcuts, roadcuts, skid trails, avalanche slopes and in areas recently exposed by retreating glaciers. Though unconstrained by the levels of soil nitrogen, Sitka alder grows better on sites rich in calcium, magnesium, phosphorus and potassium (Krajina et al., 1982). Growth rate o f Sitka alder depends on climatic and soil conditions that change with latitude and elevation. Harrington and Deal (1982) reported that Sitka alder grows on average 25-30 cm per year and after ten years can reach up to 4 m in height. The rate of growth measured as height increment decreased with increasing elevations. This plant starts to produce seeds at age 6-8 years. Seeds, extremely light and abundant, can be carried over long distances by wind and water, which helps the plant to colonize new areas (Fowells, 1965). Sitka alder is also capable of vegetative reproduction by sprouting from the stump or exposed roots, which may be important for the persistence and enlargement of established populations. Asexual reproduction has significant implications from the genetic perspective. Sitka alder plays an important role in slope stabilization, road deactivation and erosion control on steep hills. This nitrogen-fixing species improves site conditions for more valuable trees by adding nitrogen and organic matter to the soil. Bollen et al. (1969) observed the increase in carbon content in soils under Sitka alder. Binkley (1984) reported elevated levels of total and available nitrogen in Sitka alder stands. Due to its short stature, this plant has minimal importance as a competitor in comparison with other alders of coastal British Columbia, and therefore it may be more desirable for interplanting with conifers. Advantageous effects of Sitka alder on Sitka spruce (Picea sitchensis (Bong.) Carr.) and Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) grown on nitrogen poor sites are reported in the literature (Viereck and Little, 1972; Binkley, 1984). Because of its growth characteristics, this species can also be used under power and telephone lines. 2.2 Germination Seed germination is probably the first stage in plant life during which strong selection occurs since germinants are much more susceptible to unfavourable environmental conditions than are dormant or quiescent seeds. The selection acting at various stages of plant development drastically reduces the number of plants that reach maturity in comparison to the number of seeds produced. Different mechanisms in different species have evolved to prevent germination in hostile environments and stimulate germination in favourable conditions. For example, 5 stratification requirements prevent seed germination in fall or winter, heat requirements prevent germination too early in spring, phytochrome pigment systems enable some seeds to respond to light quality, quantity and duration (Kramer and Kozlowski , 1979). Deviations from the population tendency to germinate in particular conditions are often observed and widen the range of possible conditions for germination. This ability of individual seeds from a population of seeds to germinate in various conditions is itself another adaptive trait that can increase survival i f unusual circumstances occur (e.g., late frost). Since environmental conditions can vary substantially within the natural range of most species, it is not surprising that germination parameters can also vary among different populations in addition to within-population variation. Genetic control of germination attributes were found, for example, in Douglas-fir (El-Kassaby et al., 1992), Pacific silver fir (Abies amabilis (Dougl.) Forbes) (Davidson et al., 1996), Sitka spruce (Chaisurisri et al., 1992), palebark Heldreich pine (Pinus leucodermis Antoine ) (Giannini and Bellari , 1995), aspen (Populus tremula L . and P. tremuloides Michx.) (Gallo, 1985), and yellow poplar (Liriodendron tulipifera L.) (Barnett and Farmer, 1978). Germination tests are often based on seeds collected from natural stands where family structure is not preserved in the collection. In such cases, it is difficult to determine i f the germination attributes are controlled by genetics or were affected by the local environment. Environmental conditions such as temperature, moisture, nutrient availability, and even light conditions during seed development can affect subsequent seed germination (Heide et al., 1976; Nooden et al., 1985; Orozco et al., 1993). Different population environmental preconditioning has been suggested as a possible partial cause for differences in germination parameters for western hemlock (Tsuga heterophylla (Raf.) Sarg.) (Campbell and Ritland, 1982), Pacific silver fir (Davidson et al., 1996) and paper birch (Betula papyrifera Marsh.) (Bevington, 1986). From 6 the conservation point o f view it is very important to know the extent o f the genetic control in germination since early selection during nursery production favours seed/seedlings with faster germination and therefore it can reduce the genetic variability of planting stock (El-Kassaby et al., 1992; Davidson et a l , 1996). In addition, this information may be useful for ex situ genetic conservation efforts. Forms of ex situ conservation, affected by seed germination attributes, include seed storage, seed orchards, tree arboreta and plantations (Morgenstern, 1996). 2.3 Plant productivity, gas exchange and water relations There are many reports in the literature of genetically based differences evaluated at the population level in plant productivity. For example, Ager et al. (1993) determined that coastal populations o f red alder had higher growth rates than populations from less temperate inland sites and from higher elevations. Similar results were obtained in a study with black cottonwood (Populus trichocarpa Torr. & A.Gray) by Dunlap et al. (1994). Black cottonwood clones collected from climatically diverse sites were grown in a common-garden and were evaluated for various physiological characteristics. Significant differences in growth rate were observed among the populations. Variable population performance (expressed in terms of growth rate) may result from many factors. One of the important elements affecting growth rate is net photosynthesis. For example, Campbell and Rediske (1966) found that growth rate of Douglas-fir increased with increasing photosynthetic rate calculated in pgC02/min/g dry leaf material. Since the rate of CO2 assimilation changes with plant development, it is desirable to measure this attribute several times during the growing season. In a study with loblolly pine (Pinus taeda L . ) , Boltz et al. (1986) found positive correlation between the growth rate and net photosynthesis (expressed per seedling or per unit leaf area) when photosynthesis was calculated as a growing season average. 7 However, when expressed per unit area of the leaf, net photosynthesis is often not the best indicator of plant growth potential (Kramer and Kozlowski , 1979). Plant growth rate can be more dependent on the total leaf area per plant than on photosynthesis rate expressed per unit leaf area. Boltz et al. (1986) found a correlation between net photosynthesis per seedling and final plant dry weight of 0.92. On the other hand, the correlation between final dry weight and photosynthesis expressed per unit area, though statistically significant, was much lower (r = 0.46). Rates of growth of four populations of lodgepole pine (Pinus contorta Dougl. ex Loud.) were not significantly correlated with rates of photosynthesis (Sweet and Wareing, 1968). Mebrahtu and Hanover (1991) found an inverse relationship between photosynthesis and total dry weight of black locust (Robinia pseudoacacia L . ) . Results of these and many other studies indicate that the positive relationship between plant productivity and net photosynthesis may be camouflaged by factors that determine the final plant size more strongly than the rate of carbon assimilation per unit leaf area. Rate of photosynthesis (A) often depends on the concentration of CO2 in photosynthetically active tissues (Figure 1). Initially, A increases linearly with increasing C i as shown in Figure 1 (Farquhar and Sharkey, 1982). The slope of the line (A/Ci ) , often called mesophyll conductance (its inverse is called mesophyll resistance), represents limitations imposed on A by components other than stomatal conductance. Ludlow and Jarvis (1971) listed the following three components: (1) an excitation resistance important only at low light, (2) resistance to CO2 diffusion in the liquid phase from the intercellular spaces to the site o f carboxylation, (3) carboxylation resistance (or its inverse, called carboxylation efficiency). Carboxylation efficiency is proportional to the amount of the enzyme ribulose bisphosphate carboxylase-oxygenase (Rubisco) in the chloroplasts. Since the excitation resistance plays a role only at low light and since CO2 diffusive resistance in the liquid phase does not usually affect A 8 substantially (Farquhar and Sharkey, 1982) it follows that A / C i often approximates carboxylation efficiency. Because of this, and to avoid confusion with mesophyll diffusive conductance, A / C i is better referred to as carboxylation efficiency rather than mesophyll conductance. Figure 1. Rate of net photosynthesis (A) versus intercellular CO2 concentration (Ci). After Farquhar and Sharkey (1982). With C i increasing further (Figure 1), the corresponding increase in A departs from the initial linear relationship when the capacity to regenerate ribulose bisphosphate (RuP 2 ) becomes limiting. A still increases slightly with C i due to an increase in amount of RuP2 diverted from oxygenation to carboxylation, i.e. due to inhibition of photorespiration (Farquhar and Sharkey, 1982). In this region of the curve, photosynthesis is only weakly affected by stomatal 9 conductance (gs). In order to estimate stomatal limitations on photosynthesis, the linear relationship between C i and A is often assumed for all values of gs. This assumption is usually incorrect and has led to wrong conclusions of large stomatal limitations imposed on carbon assimilation rates. In fact, the limitations are usually small except for special cases in which reduced gs can cause reduction in A : treatment with abscisic acid leading to stomatal closure but not affecting assimilation capacity, reduced ambient vapour pressure, and stomatal transient behaviour in the form of overshooting or delayed response after perturbation (Farquhar and Sharkey, 1982). In a field experiment with loblolly pine (Fites and Teskey, 1988) it was shown that both mesophyll and stomatal limitations affected net photosynthesis but mesophyll processes had a greater effect than stomatal activities. The difference between net photosynthesis of two clones of Monterey pine (Pinus radiata D . Don) was related to the large differences in carboxylation efficiency observed between both clones (Bennett and Rook, 1978). The ability of plants to respond directly to changes in atmospheric vapour pressure deficit by closing or opening stomata can affect not only plant water status but also carbon assimilation rate as mentioned before. This ability has been observed in many plants (Shulze and Hal l , 1982) and can serve as an indication of adaptation to water stress. High sensitivity to vapour pressure deficit can have a negative impact on net photosynthesis as closed stomata restrict gas exchange. Such a conclusion can be reached i f a reduction in A is accompanied by a corresponding reduction in C i ; otherwise, the change in A may be due to mesophyll processes. A negative correlation between net photosynthesis and sensitivity to atmospheric humidity was indeed observed by Dang et al. (1994). In general, stomatal conductance is adjusted to minimize water loss through transpiration (E) and, at the same time, to facilitate carbon dioxide diffusion. Since both goals are mutually exclusive, stomata function to maximize unit carbon gained per unit water lost. A variable 10 defined as the ratio of the amount of CO2 fixed to the amount of water transpired, called water use efficiency ( W U E ) , brings together plant productivity and water consumption. On a short-term basis, there are two measures of W U E : (1) ratio of A to E , called instantaneous W U E (2) ratio of A to gs, called intrinsic W U E (Ehleringer et al., 1993). Since E depends on stomatal characteristics as well as atmospheric conditions (humidity gradient, temperature), intrinsic W U E is often preferred as it expresses biological functions only. However, it should be noted that since gs can respond to vapour pressure deficit, neither expression is ideal i f V P D is not held constant. They are even less useful for growth and productivity experiments since they are based on short-term measurements and do not take into account loss of carbon due to respiration and loss of water at night. Generally, plants adapted to xeric environments are expected to have high W U E , as shown in a study with black cottonwood (Bassman and Zwier, 1991). However, under extreme competition, high W U E may not be a desirable trait because it means that more water may be left available for competing vegetation. Dunlap et al. (1993) found that clones (black cottonwood) from xeric regions of Washington had lower water use efficiency than clones from mesic environments. N o differences in W U E were found among populations of red alder (Dang et al., 1994). When water supply is not a problem, high W U E may not be beneficial as was shown for field-grown wheat (Triticum aestivum L.) where it correlated negatively with the growth rate (Condon et al., 1987). M u c h like photosynthesis analysis, measurements of respiration have potential uses in assessing plant productivity and much like with photosynthesis, the success in correlating growth with respiration rates is variable. For example, Anekonda et al. (1994) found strong positive correlation between respiration and growth of coast redwood (Sequoia sempervirens (D.Don) Endl.) clones. On the other hand, negative or non-significant correlations were found for 11 populations of European silver fir (Abies alba M i l l . ) by Larsen and Mekic (1991). Total respiration has two components: maintenance respiration and growth respiration (Kozlowski and Pallardy, 1997a). Maintenance respiration is related to the biomass of l iving cells as it provides energy and compounds needed to support existing tissues. Differences among the populations in this component of respiration may translate into differences in productivity and in this case there would be an inverse relationship between respiration and growth. If, on the other hand, the differences in respiration among the populations relate to the growth processes while maintenance respiration remains the same for all populations, there would be a positive correlation between total respiration and biomass accumulation. Unfortunately, except for special circumstances (e.g. starved tissues), both maintenance respiration and growth respiration are confounded in the single measurement of carbon dioxide efflux. 2.4 Frost hardiness Species with wide distribution range growing in diverse climates often show differentiation at the population level with respect to frost hardiness. Studies with lodgepole pine, Scotch pine (Pinus sylvestris L.) and Norway spruce (Picea abies (L.) Karst.) (Rehfeldt, 1980; Eiche and Andersson, 1974; Pulkinnen, 1993, respectively) revealed positive correlations between latitude and frost hardiness of populations of these trees. Eiga and Sakai (1984) found that frost hardiness in different populations of Sakhalin fir (Abies sachalinensis Mast.) increased with altitude. What is more interesting, genetic variability with respect to frost hardiness decreased with increasing elevation. Perhaps fewer genotypes were able to survive harsh winters at higher elevations. Similar results were obtained by Will iams (1984): northern populations of red ash (Fraxinus pennsylvanica Marsh.) showed little or no significant variance in frost tolerance while southern populations were consistently variable. These findings imply a limit of 12 frost resistance that can be achieved either by artificial breeding or by natural selection. Further increase can only happen by mutation or by biotechnological techniques of gene transfer. Adaptation to local conditions can lead to the development of varieties. For example, Douglas-fir has two varieties: coastal 'menziesii' and interior 'glauca'. The coastal variety is much less hardy but more productive than the interior variety (Wright et al., 1971). There can also be differences in cold hardiness within the same variety: to avoid early frost damage, Douglas-fir 'menziesii' from the north sets buds and develops cold resistance earlier than the same variety from the south (Kramer and Kozlowski , 1979). Although the increase in population frost resistance with latitude and altitude is commonly observed, it is by no means a universal rule. Eastern cottonwood (Populus deltoides Marsh.) and black wi l low (Salix nigra Marsh.) survived exposure to at least -50° C regardless of the geographic origin of the plants (Sakai and Weiser 1973). Also , no difference in frost hardiness was observed in Scotch pine coming from low and high elevations with elevation range of 700 m (Sundblad and Andersson, 1995). It was suggested that perhaps the gene flow between the high and low altitude caused the lack of differentiation. With no barriers to gene exchange, variability in frost resistance between the populations would be low despite different climatic conditions. Studies with other species (Rehfeldt, 1988) clearly indicate that high elevation can strongly influence frost hardiness development. The ability to survive the lowest winter temperature is not the only factor limiting plant distribution. Frost hardy eastern cottonwood and black wi l low from the central and eastern regions of North America survive -80°C, yet their northern distribution is limited (Sakai and Weiser, 1973). The possible explanation for such an inconsistency is that the two species either start to grow too early in the spring or they develop frost resistance too late in the fall. Mohn and Pauley (1969) reported that races of eastern cottonwood growing in the south suffered greater 13 winter injuries than the northern races, even though both geographic races can withstand the same low temperature when frost hardy. This suggests that the time of hardiness development in eastern cottonwood plays a significant role in its survival during winter. In another example, red osier dogwood (Cornus stolonifera Michx.) from the coastal region suffered severe injury in late fall and early winter despite the fact that it survived -90° C in December (Smithberg and Weiser, 1968). From these two examples, it can be concluded that in addition to the ability to endure the lowest winter temperature, the timing of phenological events is critical to plant survival. 2.5 Repeatability 1 Phenotypic variance (Vp) is the sum of genotypic variance (VQ) and environmental variance (VE ) . The genotypic variance can be further partitioned into additive variance ( V A ) and non-additive variance: variance due to dominance (VD) and due to gene interaction (V)). The environmental variance can be divided into special environmental variance (VES) caused by temporal or localized conditions, and general environmental variance due to permanent and/or non-localized conditions (VEG)-One purpose of multiple measurements is to account for the special environmental variance. Taking several measurements of the same characteristic on the same individuals allows for the partitioning of the phenotypic variance into within individuals and between individuals components. The measurements can be repeated in time or in space. The variance within individuals is entirely due to the temporal and/or spatial differences in the environment. The variance between individuals is caused by a combination of genetic factors and non-localized environmental factors affecting individuals permanently. The ratio of the between individuals variance (VQ + V E G ) to the total phenotypic variance is called repeatability (Rp). 1 This section is based on Falconer (1989). 14 Rp = ( V G + V E G ) / Vp = ( V G + V E G ) / ( V G + V E G + V E S ) Repeatability provides an estimation of the maximum possible value of heritability in a broad sense (VQ/VP) and consequently in a narrow sense (VA/VP) . Generally, heritability is the fraction of the phenotypic variation in a particular characteristic that is caused by genetic factors. The heritability can never be greater than repeatability; however, it can be much smaller. The value of repeatability depends on the environmental conditions to which individuals are exposed, on the genetic characteristics of the population and on the measured trait itself. Two assumptions should be met before repeatability is calculated and interpreted: (1) measurements have equal variances and have the same proportions of components, (2) genetically the same character is determined by different measurements, i.e. the genetic correlation between the measurements must be one. The genetic correlation is the correlation of breeding values. The breeding value of an individual is defined as the sum of the average effects of the individual's genes. In practice, these assumptions are often not fully met. Repeatability can also be used to estimate improvement in accuracy that results from taking several measurements. The lower the repeatability, the greater the improvement. Taking more than three measurements usually does not substantially improve accuracy. 15 3.0 M A T E R I A L S A N D M E T H O D S 3.1 Plant material Sitka alder seeds were collected by the Research Branch of British Columbia Ministry of Forests in fall of 1994 from 10 sites along coastal British Columbia. In general, each collection was made from at least 10 well distributed individual trees a minimum of 10 m apart. About 2 liters of seed catkins were collected from each site. The seeds were sown at 3 seeds per cavity in styroblocks (PSB313B®, cavity volume 65 mL) on Apr i l 6th, 1995 in the Saanich Forestry Centre in Saanichton (latitude 48° 35', longitude 123° 24', elevation 50 m). On October 30th, 1995 eighty seedlings per population (except for the Squamish population which had 54 seedlings) were potted into plastic containers (container volume 2650 mL) for a total of 774 seedlings. In addition to the 10 coastal populations, 22 new seed sources were added to the study the following year. The seeds from all 32 populations were sown in A p r i l 1996. A total o f 1958 seedlings (39 to 80 per site) were potted in October 1996. The final number of studied populations was 29 (Table 1 and Figure 2) as some seed sources did not germinate well . Not all of the 29 populations were used in every experiment due to the limited number of seedlings from some locations. It should be noted that the collection included one population from northern California (Areata); the remaining 28 seed sources came from British Columbia. 16 Table 1. Population location and symbol. POPULATION S Y M B O L L A T I T U D E L O N G I T U D E E L E V A T I O N (m) DISTANCE F R O M T H E C O A S T (km) Adam's Plateau A D 51° 06' 119° 33' 1280 330 Areata A R 40° 54' 124° 05' 1500 2 Bella Coola B E 52° 19' 126° 46' 800 90 Bute Inlet B U 50° 56' 125° 05' 960 76 Chine Nose Summit C H 54° 27' 126° 08' 1430 260 Cold Creek C O 50° 49' 120° 07' 1310 290 Cranbrook CR 49° 35' 117° 05' 640 420 Cypress Park C Y 49° 22' 123° 12' 980 6 Dease Lake D E 58° 45' 130° 03' 869 250 Glena Bay G L 50° 35' 117° 52' 607 400 Golden G O 51° 30' 117° 20' Unknown 490 Green Mountain G R 49° 03' 124° 21' 1100 20 Hemlock Valley H E 49° 23' 121° 56' 1013 80 Hope Slide HO 49° 16' 121° 15' 762 120 Kimsquit River K I M 52° 53' 127° 10' 800 110 Kitlope River KIT 53° 03' 127° 36' 820 100 Knight Inlet K N 51° 06' 125° 48' 790 70 McKay Lake M C 49° 45' 125° 17' 914 10 McKendrick Pass M K 54° 50' 126° 45' 1190 250 Owikeno Lake OW 51° 34' 126° 31' 1080 84 Phoenix PH 49° 05' 118° 35' 1219 310 Powell River PO 49° 59' 124° 39' 660 10 Roberts Lake ROB 50° 13' 125° 30' 366 10 Roger's Pass R O G 51° 19' 117° 34' 1219 460 Sproat Lake SP 49° 18' 125° 04' 640 30 Stikine River ST 58° 00' 130° 02' 915 210 Squamish SQ 49° 39' 123° 02' 785 24 Valemount V L 52° 50' 119° 15' 850 490 Vanderhoof V N 53° 56' 123° 49' 1287 370 Figure 2. Map of studied populations of Sitka alder in British Columbia. Part of British Columbia where Sitka alder does not grow is delineated in gray. 18 3.2 Germination Seed germination tests were conducted in the Saanich Forestry Centre prior to the seed sowing in March 1995 and March 1996. Four replications of at least 100 seeds each from each of the 10 coastal populations in 1995 and from the additional 19 populations in 1996 were used in the standard germination tests. The seeds were placed on moist filter paper and kept in warm temperatures (20 to 30 °C) for 21 days. The seeds were not stratified. After seven days germinants were counted every day and assessed according to I S T A (International Seed Testing Association, 1985) rules. The following parameters were determined: G C (germination capacity), R50 and R ' 5 o (germination rates), P V (peak value), G V (germination value) and G U (germination uniformity). G C is the percentage of total germinated seeds; R50 is the number of days it takes to germinate 50% of the total seeds (Ching 1959); R'50 is the number of days it takes to germinate 50%) of the viable (and non-dormant) seeds (Thomson and El-Kassaby, 1993). In order to estimate P V , the accumulated number of germinants was divided daily by the number of corresponding days. The maximum value obtained is called P V and it represents the mean daily germination of the most vigorous seeds (Czabator 1962). G V is computed by multiplying P V by M D G (mean daily germination) and it represents germination speed and/or germination completeness (Czabator 1962). A parameter describing germination uniformity (GU) was devised as a sum of the highest daily germination rates for three consecutive days and it was expressed as a proportion of the total viable seeds. 3.3 Frost hardiness and bud burst Depth and induction of frost hardiness was measured separately in one- and two-year old plants. One-year old plants were represented by 27 to 29 populations while two-year old plants 19 were represented by the 10 initial coastal populations. A total of 15 tests (8 for two-year old and 7 for one-year old plants) were conducted from 01 October, 1996 to 09 March, 1997. Sample size for each trial ranged from five to 10 plants per population. Frost hardiness was evaluated by the measurement of electrolyte leakage from segments of stems exposed to sub-freezing temperatures following, with small modifications, Si lvia L'Hirondelle (personal communication, 1996). The conductivity of the tissue solution was measured with a digital conductivity meter (model 1481-60, Cole Parmer) with a gold plated dip cell (model 1481-62, Cole Parmer). Roots, buds and leaves were cut off and the stems were washed in deionized water before they were cut into 5 mm long sections, 16 sections per plant. Equal numbers of sections (four) were put in four 7-mL polyethylene scintillation vials and 0.2 m L of deionized water was added. One vial per plant was kept in the refrigerator at 4°C as control. The remaining three samples from each plant were placed in a programmable freezer (Forma Scientific) and each was exposed to a different subfreezing temperature. The temperatures were selected based on the results of previous tests and preliminary trials and varied from -6 to -50°C depending on the level of frost resistance developed by Sitka alder at the particular time. Two blank vials were included for control and each treatment temperature. One test was conducted to determine frost hardiness repeatability. In that case, all three vials per plant were exposed to the same temperature. The samples were cooled at the rate of 4°C per hour and kept in each test temperature for one hour. After that time, the vials were removed from the freezer and thawed for at least 2 hours at 4°C. Next, 3.3 m L of deionized water was added to the control and test vials and the samples were allowed to equilibrate at room temperature for 18 hours. Following the conductivity measurements the tissues were kil led in a water bath (90 °C for one hour) and left for 18 hours before the second conductivity measurement. Relative conductivity (RC) and frost injury index (FII) were calculated as below. Population mean lethal temperature that kil ls 50% of the tissues (LT50) was 20 determined by regressing population mean FII on treatment temperatures. Since some of the plants had FII above or below 50% for all three test temperatures, LT 5 o was not calculated for individual plants. The approximate date by which each population developed frost resistance to -18°C (expressed as LT50 (DATELTSO(-I8))) was obtained by regression. R C = (C1 -B1 )* 100/(C2-B2) FII= (RC t -RC c ) / (100-RC c ) where: C I : conductivity following low temperature treatment (mho) C2: conductivity of kil led tissues (mho) B l : mean conductivity of the two blanks after exposure to low temperature B2 : mean conductivity of the two blanks after exposure to 90 °C R C t = relative conductivity of sample exposed to treatment temperature (%) R C C = relative conductivity of control sample (%) FII = frost injury index (%) Date of population bud burst was determined by five inspections of all 1958 seedlings from 28 populations between 09 March and 15 Apr i l 1997. The emergence of the first leaf from the terminal bud was used as a criterion for bud burst. Seedlings starting to grow were counted and expressed as a percentage of the total number of plants in each population. Date when 50% of the population seedlings flushed was determined by regression. The dates were converted to the number of days starting from 15 March (DAYS5o%budbreak)-3.4 Gas exchange A total of 400, two-year-old plants from 10 coastal populations were randomly selected in July, 1996. Exchange rates of carbon dioxide and water vapour were determined using an open 21 gas exchange system (model L C A 3 , The Analytical Development Company) with a Parkinson broadleaf chamber (model P L C 3 , A D C ) . The carbon dioxide concentration of air entering the leaf cuvette was controlled at 363.6 ± 17.5 p L / L (mean ± 3SD); relative humidity (RH) was controlled at 45.9 + 2.4%. The measurements were made in sunlight with photosynthetically active radiation exceeding 1000 pmol/m 2/s of photon flux density. To minimize the instrumental error, gas exchange of each measured leaf was recorded twice and the average of these measurements was used in the analyses. Plants were measured between 8-14 July and those measured on the same day constituted one block for statistical analysis. Based on carbon dioxide and water vapour exchange rates the following parameters were estimated as in von Caemmerer and Farquhar (1981): transpiration rate (E, mmol H 2 0 m " 2 s"1), stomatal conductivity (gs, mmol H 2 0 m ~ 2 s"1), net photosynthesis (A, pmol C 0 2 m" 2 s"1) and ratio of intercellular to ambient C 0 2 concentration (Ci/Ca). Photosynthetic intrinsic water use efficiency ( W U E ) was calculated as A to gs ratio. Assuming negligible impact of C 0 2 from photorespiration on intercellular C 0 2 concentration, carboxylation efficiency (gm, mol C 0 2 m" 2 s"1) was obtained by dividing net photosynthesis by intercellular C 0 2 concentration (Ludlow and Jarvis, 1971). In July, 1997 a total of 700 two-year-old plants from 28 populations (25 seedlings per population) were randomly selected for gas exchange measurements. The plants were measured between 19-25 July. The same procedure was employed as explained for tests conducted in July 1996 and the same parameters were estimated. However, an artificial light source (35 W Eye Dichro-Cool halogen lamp, Iwasaki Electric Co.) was used to ensure the delivery of 1200 pmol/m 2/s photon flux density. Measurements were taken on sunny days ensuring that the plants were induced to high light. The concentration of carbon dioxide of air entering the leaf cuvette was controlled at 329.3 ± 7.8 p L / L ; R H was controlled at 45.4 ± 4.0%. Stomatal sensitivity to water vapour pressure deficit ( V P D s e n s ) was calculated by 22 measuring transpiration rates and stomatal conductance under low and high relative air humidity. A total of 164 plants from 10 coastal populations were selected for this experiment. The plants were measured between 23-27 July, 1996. Since the relationship between stomatal resistance and vapour pressure deficit is linear (Dang et al., 1994), the low and high R H values were allowed to vary to some extent from plant to plant but on average, low R H was 30.6 ± 9.6% and high was 51.8 ± 10.8%. C 0 2 was controlled at 361.3 ± 15 p L / L . In September, 1997 a total of 112 plants from 28 populations (four per population) were selected to measure V P D s e n s using the same procedure. The average low R H was 18.8 ± 15.3% while high was 40.9 ± 18.3%. C 0 2 was controlled at 329.6 ±12 p L / L . A total of 150 plants from 10 coastal populations were sampled July 24-26, 1996 to examine their dark respiration rates (R). The respiration was measured indoors between 7 and 10 p.m. on the same plants which had their photosynthetic rates examined during the day. R H was controlled at 55.4 ± 6.6%, and C 0 2 was 375.1 ± 42 p L / L . During the following year (1997) a total of 280 plants (10 per population) from 28 populations was sampled from 19-24 July. R H was controlled at 45.8 ± 3.6%, and C 0 2 was 331.7 ± 14.1 p L / L . To estimate repeatability of the gas exchange characteristics, spatial variations were estimated by taking measurements on three leaves from every tested plant. A total of 84 plants form 28 populations (three per population) were measured using an open gas exchange system (model L C A 4, A D C ) . The measurements were taken between 12-13 August, 1997. R H was controlled at 44.8 ± 6.6%, and C 0 2 was 329.9 ± 4.5 p L / L . 3.5 Quantitative attributes Height was measured on all 774 seedlings from the 10 initial coastal populations on 23 March, 1996. Thirty six percent of these plants eventually died in late spring 1996, possibly due 23 to frost damage to the root systems. Height of the remaining 494 plants was measured on 26 June and 27 July 1996. In September, 1996, 248 plants were subject to destructive sampling. The following biomass allocation parameters and plant architecture characteristics were measured: total leaf area per plant ( L A ) , shoot dry weight (SDW), root dry weight (RDW) , main stem height and diameter (H, C A L ) , number of primary and secondary stems or branches within the first 6 cm from the shoot-root transition zone that had diameter greater than 4 mm (STEMS) and length together with diameter of the largest three stems or branches, starting with the biggest one (SI, S2, S3 and C A L 1 , C A L 2 , C A L 3 ) . Plant architecture index (PAI) was calculated as (S2+S3)/S1+(CAL2+CAL3)/CAL1. The maximum value of P A I is 4. The higher the number, the more "bushlike" the plant. Height of a total of 700 plants from 28 populations (25 per population) was measured on 30 March, 06 May, 04 June, 09 July, 12 August, 04 September and 25 September 1997. Absolute height growth rates (cm day"1) were calculated for each month of the growing season. In November, 1997 these plants were destructively sampled. The same procedure was followed as in 1996 except that L A was not measured. 3.6(a) Statistical analysis A l l variables were subject to analysis of variance ( A N O V A ) or covariance ( A N C O V A ) , multiple linear regression and simple product-moment correlation analysis. Normality was checked using the Kolmogorov-Smirov test (SigmaStat, 1994). Homoscedasticity was checked by the Levene Median test for A N O V A and by the Spearman rank correlation for regression analysis (SigmaStat, 1994). Where needed, appropriate transformations were found to satisfy the assumptions. In addition to the univariate procedures, multivariate analyses were utilized: cluster analysis on frost hardiness and gas exchange data, factor analysis on germination data and 24 canonical correlations on all data sets except frost hardiness. The G L M procedure of S A S was employed for A N O V A and A N C O V A along with the V A R C O M P procedure using the restricted maximum-likelihood method of estimation ( R E M L ) to calculate variances (SAS, 1988). The following transformations of variables were used to meet the assumptions of normality and constant variance: R'5oT=l-(l/(R' 5o+l)); G C T = arcsin(GC/100); GVT=sqrt(GV+0.5); GUT=ln(GU); A T = A 1 5 ; gsT=gs 0- 2 5; Ci/CaT=(Ci/Ca) 6 - 1 ; WUET=log(WUE) , VPDSensT=(log(VPDse„s+2))"1; SDWT=sqrt(SDW), R D WT=sqrt(RD W); TDWT=sqrt(TDW), GRJ=sqrt(GRJ). Significant covariates included lt for gs, W U E , C i / C a and E ; and time and time for A and gm. N o significant covariates were found for R and VPD s e ns-Multiple linear regression, with latitude ( L A T ) , longitude ( L O N G ) , elevation ( E L E V ) and distance from the coast ( C O A S T ) as independent variables and population means as dependent variables, was employed using the R E G procedure (SAS, 1988). The forward stepwise procedure with a 0.15 significance level for entry into the model was used to eliminate variables not contributing to the variance of the response variables. Pearson product-moment correlations were calculated for population means. The probabilities of the observed values for the coefficients of correlation were adjusted for the number of pairs of variables in each test using Bonferonni's procedure ( S Y S T A T , 1996). Canonical correlation based on population means was calculated for germination, gas exchange and biomass parameters. Canonical correlation analysis estimates the relationship between predictor (independent) variables and criterion (dependent) variables simultaneously (Dil lon and Goldstein 1984). Geographic variables ( L A T , L O N G , E L E V , C O A S T ) were used in the C A N C O R R procedure (SAS, 1988) as predictor variables. The purpose of canonical analysis was: (1) to determine how much of the total variance of the dependent variables could be explained by the linear combination of the location variables, i.e. by their canonical variates, (2) 25 to obtain cross-loadings which represent correlation between individual variables of one set and a canonical variate of the other set, (3) to obtain as part of the C A N C O R R output simple correlations among and between location variables and response variables. Two hierarchical cluster analyses were conducted to group similar populations together based on frost hardiness data (first analysis) and gas exchange-related data (second analysis). Population least-squares means were subjected to a clustering procedure using the complete linkage algorithm for gas exchange variables and the average linkage algorithm for frost hardiness variables ( S Y S T A T , 1996). In both methods, the Euclidean distance was used as a measure of distance between the clusters. In the complete linkage method, the distance is defined as the furthest distance between pairs of points in the two clusters; in the average linkage method the distance is defined as the average distance between all pairs of points (Dillon and Goldstein 1984). Because of large scale differences among the gas exchange variables, these variables were standardized before the analysis by subtracting the sample mean from each value and dividing the difference by the standard deviation. A l l variables related to germination were also subjected to factor analysis ( S Y S T A T , 1996) to find groups of interrelated variables and underlying common factors. Factors were extracted by principal component analysis using the correlation matrix as an input (i.e. standardized variables). Principal components with "eigenvalues" greater than one were retained. Eigenvalues o f the correlation matrix represent the amount o f common variance extracted by a principal component (Dil lon and Goldstein, 1984). The obtained factors were subject to orthogonal rotation using the "varimax" method which maximizes the variation o f the squared factor loadings for each factor (Dil lon and Goldstein, 1984). A s a result, the number of variables that have high loadings on each factor is minimized thus simplifying the interpretation of the factors. For orthogonal factors, loadings are simple correlations between a factor and variables. 26 3.6 (b). Analysis of variance models and experimental design A completely randomized design was used to evaluate parameters related to germination, frost hardiness and biomass allocation. The data were subjected to one-way analysis of variance according to the general linear model: Yjj = p + tj + e ( i ) j p= common mean Tj= population effect; random j = 1 k k = 2 7 - 2 9 , number of populations £(i)j= experimental error i = l n n = sample size per population; n = 4 , 5 - 1 0 and 2 5 for germination, frost hardiness and biomass parameters in the 1 9 9 7 experiment, respectively. Table 2 . One-way analysis of variance; tp = intraclass correlation coefficient. Source D . F . Mean Squares Components variance of F test tp Population (P) k - 1 M S P 2 2 a EE + na P M S P / M S E E o-2p/(rj 2p+a 2 E E) Experimental k(n - 1 ) M S E E O"2EE Error (EE) Total k n - 1 A randomized complete block design was used to evaluate parameters related to gas exchange. The data were subjected to two-way analysis of variance or covariance according to the general linear model: Yjji = p + Tj + Pi + Sy + co(ij)i + covariates u= common mean Tj= population effect; random j = l k k = 28, number populations Pi = block (test date) effect, random i = l n n = number of blocks; n = 2 for V P D s e n s , 6-7 for all other parameters Sy = experimental error (B(ij)i = sampling error 1=1 P p = sample size per population per block; p = 1-5 depending on test date Table 3. Two-way analysis of variance; tp = intraclass correlation coefficient. Source D . F. Mean Squares Components of variance F test tp Population (P) k - 1 M S P 2 2 a SE + po" EE + M S P / MSEE a 2 P / ( a 2 S E + pna 2p 2 2 O" EE + a p) Block (B) n - l M S B 0"2SE + kpCT2B+ M S B / M S E E PO-2EE Experimental (k - l)(n - 1) M S E E 2 , 2 O" SE + PO" EE M S E E / M S S E Error (EE) Sampling Error k n ( p - l ) M S S E 0"2SE (SE) Total kpn-1 28 Repeatability of frost hardiness and gas exchange variables (Becker, 1984) was calculated based on the following general linear model: Ykm = p + a k + e k m p = common mean ak = effect of the k-th individual ekm = the environmental deviation of m-th measurement within an individual Table 4. Analysis of variance table for repeatability. Source D.F . M S Components of variance Between N - l M S B a 2 E + k io - 2 B Individuals (B) -Between measurements, N ( M - l ) M S E a 2 E within individuals (E) N = number of individuals; 155 for frost hardiness (5 per population, 31 populations), 84 for gas exchange (3 per population, 28 populations) M = number of measurements per individual (3) k i = M for equal number of measurements for each individual a E = represents special environmental variance (VES) a B = represents general environmental variance (VEG) and genetic variance (VG) Estimated (est)rj 2 E = M S E Estimated (est) a 2 B = ( M S B - M S E ) / k i Repeatability Rp = (est) CT2B / ((est) CT2B + (est)a 2 £) 29 4.0 R E S U L T S 4.1 Germination Statistically highly significant differences (P<0.01) among the populations were detected for every estimated germination parameter (Appendix 1). Germination attributes of the 27 populations are depicted in Figure 3. G C varied from 11.8 % to 87.2 %. Since many populations had G C below 50%, germination speed expressed as R50 was not calculated. P V ranged from 0.65 to 8.5 germinants per day while G V ranged from 0.39 (low germination completeness and/or speed) to 35.1. Speed of germination, expressed as R'50 varied from 6.7 to 13.9 days and was not significantly correlated with G C (r = -0.27, P>0.1). Populations also differed in G U as the percentage of viable seeds germinated within three consecutive days ranged from 39.5 to 92.9%. Fast germinating populations were also more uniform in germination (r = -0.88 between G U and R ' 5 0 ) . Germination parameters: G C , P V and G V were all highly correlated (r values >0.93). In this case, it can be argued that they all in fact approximate one parameter: germination completeness. Due to thelarge differences in G C , P V depends more on the germination capacity of the most vigorous seeds than on speed of germination (Figure 3). Populations which germinated quickly but had low G C , also had low P V . G V is a parameter that combines both P V and G C and therefore it also measures germination completeness. Strong dependencies among the germination parameters and possible redundancies in the data were also confirmed in the factor analysis. Out of five principal components, the first two that explain 95.6% of the total variance in germination parameters were retained (Appendix 2). Simple structure was achieved after the varimax rotation as the original variables loaded highly on either one of the two factors (Figure 4): P V , G C and G V on factor 1 (r = 0.95, 0.98, 0.96, 30 respectively), R'50 and G U on factor 2 (r = -0.94 and 0.95). It is apparent that factor 1 measures germination completeness since the germination speed or uniformity are not related to it. Factor 2 represents both germination speed and uniformity, as populations with faster germination germinated more uniformly. Simple correlation analysis and factor analysis revealed that speed of germination and germination completeness are independent in the studied Sitka alder populations. G C , G V , and P V are closely related to each other and all are approximately orthogonal to R'50 (Figure 4). Patterns of variations Based on linear regression analysis, G U did not show any relationship with population location. Very weak north-south trends were discovered for the remaining germination parameters (data not presented). The strongest correlation was found for R'50 (high latitude populations had higher speed of germination) but latitude could explain only 28% of the variation (P < 0.01). Canonical correlations detected the same weak relationship between germination parameters and geography (Appendix 3). The probability of the observed value for W i l k s ' Lambda was only 0.04. W i l k s ' Lambda is a multivariate statistic testing the null hypothesis that all canonical correlations are zero (SAS, 1988). Only the first geographic canonical variable ( L O C A T 1 ) was significant and it explained 22% of the variance of the germination variables. L O C A T 1 was related very strongly to latitude (r = 0.91). R'50 had the highest loading of all variables on L O C A T 1 (r = -0.64). 31 • 95 90 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 D A Y S S I N C E P L A N T I N G Figure 3. Cumulative germination of 27 Sitka alder populations during the standard germination tests. F A C T 0 R ( 1 ) Figure 4. Factor loadings plot for germination parameters. 33 4.2 Frost hardiness There were significant differences in frost injury index (FII) among the populations at all test dates except for the last experiment on 09 March (Figure 5). It is apparent that the largest differences among the populations were found in fall, during the development of frost hardiness. The least frost hardy population reached a LT50 of -18°C 44 days later than the most frost hardy population (Figure 6). The population from near Vanderhoof (VN) was frost hardy to -18°C on 19 October, while the high elevation plants from northern California ( A R ) did not reach this level of hardiness until the beginning of December. The populations also differed significantly in maximum cold hardiness. For example, mean LT50 calculated for 20 January ranged from -50°C to -20°C (Figure 7). The differences in maximum depth of frost resistance were less pronounced than were found for the timing of frost hardiness development. On 11 November, 65.4% of the total variation in FII was explained by the differences between the populations, while on 15 December 33.0 % of the total variance was due to the population effect (Table 5). FII variances were calculated based on data from test temperatures where population differences were most pronounced. In general, populations that started to develop frost hardiness earlier were more frost hardy in the middle of winter, but the relationship was not strong (r = -0.6, P<0.05 between January LT50 and DATE LT50(-i8))-Maximum possible heritability of frost hardiness was estimated at 0.89 using the multiple measurements method. 34 Table 5. Components of variance (a 2) for population FII evaluated at five dates. Data are presented for the test temperature which produced the largest differences among the populations at each date. D A T E a 2 P „2 ERROR (a2P/rj2TOT)-100 10 October 83.3 117.6 41.5 % 11 November 217.6 116.5 65.4 % 15 December 64.1 108.3 33.2 % 20 January 46.9 60.6 43.6 % 09 March 6.7 154.1 4.2 % Large differences among populations in time of bud break were also found but they should be regarded only as an indication of population differentiation with respect to this trait. N o significance levels are associated with the differences since time of bud break was not determined for individual plants but was only estimated for each population based on percentage of plants with flushed buds at a particular date. The first population started to grow on 15 March (ST, northern British Columbia) and the last one 25 days later ( A R , northern California, high elevation). Differences among British Columbia populations only were as large as 23 days. Patterns of variations Multiple linear regression based on population means detected a substantial clinal trend for population cold hardiness estimated during the time of hardiness development (Appendix 4). FII evaluated at -25 °C on 11 November was significantly correlated with latitude and the distance from the sea (R 2 = 0.62) according to the equation: FII = 177.05 - 0.0388 L A T (') -0.039 COAST(km) , (P<0.01). Most of the explained variation was due to latitude (partial r 2 = 0.48, P<0.01). Only a weak geographic trend was found for population maximum frost hardiness which increased with latitude (r 2 =0.31, P O . 0 1 between January LT50 and L A T ) . Cluster analysis, based on population mean FII of all tests, produced four distinguishable groups (Figure 8). Coastal southern populations were clustered with populations from the south interior wet belt. Northern interior populations were grouped with some southern populations 35 from dry interior regions. V N , the most frost hardy population and A R , the least frost hardy population were clearly very different from any other seed sources. A significant geographic trend was found for spring bud break (R 2 = 0.43). Northern and interior seed sources had a tendency to flush earlier (partial r = 0.35, P O . 0 1 with L A T and r = 0.08, P O . l with C O A S T ) . Some interior populations started to grow as late as coastal populations causing poor correlation with C O A S T . In addition, there were some examples of southern populations breaking bud relatively early (e.g. C Y ) . 36 90 T : 1 1 1 1 1 1 1 1 1 1 —T 1 1 I I I I ' ' I I . I I 15-Oct-96 l l -Nov-96 15-Dec-96 20-Jan-97 09-Mar-97 DATE Figure 5. Change in mean frost injury index (%) of individual populations evaluated at - 18°C from 15 October, 1996 to 09 March, 1997. 37 O H < OH o PH A R R O G O R G G R P O M C H E SP B U H O M K V A A D C H G L D E K N C O C R P H ST V N i i - j — i — i — i — i — i — i — r 21-Oct-96 28-Oct-96 04-Nov-96 l l - N o v - 9 6 18-Nov-9625-Nov-96 02-Dec-96 DATE OF POPULATION L T 5 0 OF -18°C Figure 6. Differences among the populations in frost hardiness development based on dates when each population attained resistance to -18°C. Frost resistance was measured as population mean LT50. 38 -50 1 F K V N DE HE GR KIT CH V A BU ST CR CO ^ M K g HO § PH £ AD ROG PO GO K N SP MC ROB AR CY -40 -30 -20 L T 5 0 O N 20 JANUARY f°Q -10 0 Figure 7. Mean population L T 5 0 (°C) on 20 January, 1997. 39 Cluster Tree V N -ST -P H -CR ' c o -M K -V L -C H D E KIT-H E -ROB GO PO ROG M C GR G L B U A D SP HO K N -A R 0 10 15 20 25 DISTANCES Figure 8. Cluster analysis dendrogram of population grouping based on frost hardiness. Population FII means of all tests and temperatures were used as input variables. 40 4.3 Gas exchange parameters The univariate analysis of covariance of the gas exchange variables revealed highly significant differences (PO.01) among the populations in gs, gm, W U E , E , A , R, and C i / C a (Appendix 5). A s much as 25% (for A ) of the total variance was due to the population effect (Appendix 5). Overall mean, standard deviation, range and relative range for population means of these variables are presented in Table 6. N o differences were detected for V P D s e n s ; however, the power of this test was low due to the small sample size (four plants per population) rendering the analysis less sensitive. Among the coastal populations only (1996 experiment) there were no differences in E , gs, W U E and V P D s e n s -Interdependence of the gas exchange parameters was expressed in high correlations for some pairs o f variables: gs and E , A and gm, gs and A (r = 0.91, 0.97 and 0.92, respectively). A and gs are both linked through the C i value and, as a result, were also positively correlated with each other as well as with C i /Ca . C i / C a was correlated with W U E (r = -0.99), gs (r = 0.89) and A (r = 0.70, P<0.01). Dark respiration rates did not correlate significantly with any gas exchange parameter. On average, 80%. of the total genetic variance of gas exchange variables was found within the populations (tp in Table 6). Repeatability values ranged from 0.52 to 0.84 (Table 6). 2 The results presented pertain to the 1997 experiment. The 1996 experiment is referred to but the results are not shown in detail. Al l provenances but one from the 1996 experiment were included in the 1997 experiment. 41 Table 6. Overall mean, standard deviation (SD), means' range and relative range (RR = range 100/mean), intraclass correlation coefficients (tp) and repeatabilities (Rp) of variables related to gas exchange. Overall mean and standard deviation were calculated using individual observations. Range and relative range were calculated for population means. Variable unit Overall mean SD Range of means R R (%) Rp tp E mmol H 2 O m"V 3.25 0.72 2.90-3.97 32.9 0.84 0.19 A pmol C O 2 m 'V 1 11.66 2.54 10.35- 39.0 0.72 0.31 14.90 gs mmol H 2 O m 'V 1 645.1 406.8 406.0- 106.3 0.77 0.20 1091.6 gm mol C 0 2 m 'V 1 0.0468 0.01 0.0407- 38.7 0.52 0.26 0.0588 C i / C a 0.835 0.054 0.821-0.876 6.6 0.52 0.12 W U E pmol C 0 2 / m m o l H 2 O 0.023 0.011 0.0135- 50.9 0.64 0.13 0.0252 R pmol C 0 2 m 'V 1 2.19 0.476 1.74- 38.4 N / A 0.13 2.58 Patterns of variations A significant geographic trend was detected for the population instantaneous rate of photosynthesis (Appendix 6). Mean A increased with C O A S T , E L E V and L A T according to the equation: A = const + 0.0035COAST + 0.0013ELEV + 0.0013LAT (R 2 = 0.52, P<0.15 for each model parameter). Most o f the explained variation was due to higher photosynthetic rates of the interior populations compared to the coastal populations (partial r for C O A S T = 0.41, P<0.01). Variations in A over the range of temperatures for coastal and interior populations are presented in Figure 9. The interior populations had consistently higher A than coastal seed sources but the decrease in A with increasing temperatures was the same in both groups. Since C i / C a was also decreasing with rising temperatures (Figure 10), the temperature increase did not affect A directly but rather indirectly by increasing V P D which reduced stomatal conductance. There were no significant differences between coastal and interior C i / C a evaluated at the same temperature 42 except for 25.5 °C where the interior C i / C a was statistically higher than coastal C i /Ca . The higher A over the same range of temperatures in the interior populations, compared to the coastal seed sources, was due to a higher carboxylation efficiency and higher stomatal conductance: the interior populations had greater A compared to the coastal populations at the same C i / C a and gs as shown in Figures 11 and 12, respectively, and higher gs over the same range of temperatures (Figure 13). Greater gs allowed the interior plants to maintain the same or slightly higher C i / C a compared to the coastal plants even though the interior plants had higher carboxylation efficiency. Less strong, but significant geographic trends were also found for all other gas exchange related variables except V P D s e n s (Appendix 6). For example, interior populations had higher gm (r 2 = 0.39) as was also shown above; gs increased with C O A S T (partial r 2 = 0.25, P<0.01) and with E L E V (partial r 2 = 0.08, P<0.1). Variations in gs for interior and coastal populations over the range of temperatures are presented in Figure 13. Similar to the geographic pattern of variation in A , the interior populations had higher gs compared to the coastal populations. Stomatal conductance decreased markedly with increasing V P D (Figure 14). A t lower V P D interior populations seem to have higher gs compared to the coastal plants as there were significant differences at V P D o f 0.65 and 0.75 kPa. Interior populations maintained greater gs than coastal populations at the same temperatures (Figure 13) and the same R H of air entering the cuvette. Although gs was decreasing with increasing temperatures above 26 °C, the stomata clearly responded to changes in V P D rather than temperature (Figures 13 A and B) . Stomatal conductance remained unchanged over the range o f temperatures up to 26 °C because the increased transpiration rate (Figure 13 B) maintained the same V P D even though the temperatures were increasing. Once the transpiration rate reached a plateau, gs decreased rapidly in response to the increasing V P D with 43 increasing temperatures (Figure 13 A ) . Cluster analysis based on all gas exchange variables demonstrated good coast versus interior division in Sitka alder populations rather than latitudinal or altitudinal differentiation (Figure 15). Wi th some exceptions, coastal and interior populations were grouped into two different groups. Exceptions included grouping of two interior populations, V L and D E with coastal populations and two coastal populations, A R and C Y , with interior populations. However, based on photosynthetic rate only, A R and C Y were correctly included with coastal populations (data not presented). A s already presented, multiple linear regression analysis detected geographic trends in all variables related to gas exchange except V P D s e n s - It also showed some differentiation in A along latitude and elevation in addition to the strong correlation with C O A S T . In contrast, canonical correlation analysis (Appendix 7) did not find any significant geographic trend other than coast -interior differentiation in A and gm. The first canonical geographic variable ( L O C A T 1 ) was related mainly to elevation (r = 0.66, P<0.01). However, no significant correlation was found between L O C A T 1 and any gas exchange variable. The second canonical geographic variable ( L O C A T 2 ) represents a measure of continentality, as both C O A S T and L O N G were correlated with it (r = 0.99 and -0.69, respectively). The remaining two canonical correlations were not significant. The second canonical correlation explained the largest proportion of the gas exchange variables variance (20%) from all constructed canonical correlations. A and gm were the only variables that loaded significantly (P<0.05) on L O C A T 2 . 44 Figure 9. Mean rate of net photosynthesis (A) of interior and coastal populations versus temperature. Each mean was calculated based on temperature interval of 1 degree. Error bars represent standard error of the mean; n.s. = non-significant difference (P>0.05). 45 0.92 -j 0.90 -0.76 -0.74 -I 1 1 1 — ' 1 20 22 24 26 28 30 32 TEMPERATURE (°C) Figure 10. Mean C i / C a of interior and coastal populations versus temperature. Each mean was calculated based on temperature interval of 1 degree. Error bars represent standard error of the mean; s.d. = significant difference (PO.05) . 46 Figure 11. Mean rate of net photosynthesis (A) of interior and coastal populations versus C i /Ca . Each mean was calculated for the same temperature intervals as in Figure 9. Error bars represent confidence intervals of the means (95%). 47 Figure 12. Mean rate of net photosynthesis (A) of interior and coastal populations versus stomatal conductance (gs). Each mean was calculated for the same temperature intervals as in Figure 9. Error bars represent confidence intervals (95%) of the means. Hyperbolic curves (y=ax/(b+x)) were fitted for both data sets. 48 0 -I 1 1 1 1 1 1 20 22 24 26 28 30 32 TEMPERATURE (°C) Figure 13. Mean stomatal conductance (gs) of interior and coastal populations. Error bars represent standard error of the mean; n.s.=non-significant difference (P>0.05). Each mean was calculated based on temperature interval of 1 degree. (A) vapour pressure deficit ( V P D ) at the leaf surface and (B) transpiration (E) of interior and coastal populations versus temperature. 49 1600 1400 1200 1000 5 800 So 600 H 400 200 Figure 14. Mean stomatal conductance (gs) of interior and coastal populations versus vapour pressure deficit (VPD) . Each mean was calculated based on V P D intervals of 0.1 kPa. Error bars represent standard error of the mean; s.d =significant difference (P<0.05). 50 Cluster Tree G O ST GL PH ROG AD VN CH C O MK CR AR CY KIT BU DE HE HO PO KN SP GR OW ROB BE VL KIM MC o Distances Figure 15. Cluster analysis dendrogram of population grouping based on gas exchange data. Population means of all gas exchange related variables were used as input variables. 51 4.4 Quantitative attributes3 Statistically highly significant differences (P<0.01) among the populations were found for all variables related to plant morphology (Appendix 8). A s much as 61% (for H I ) of the total variance was explained by population effects. Table 7 includes the overall mean, standard deviation, range and relative range for population means. On average, 59% of the genetic variance of variables related to biomass was found within the populations. Significant differences (P<0.01) were also found among the coastal populations only (1996 experiment) for all biomass related variables including L A which was not measured in 1997. Changes in mean population height and height growth rates during the second growing season are presented in Figures 16 and 17, respectively. It is apparent that the height growth in the first and in the second season are not necessarily related. For example, H E was the tallest population in 1996; in 1997 it was surpassed by 10 other populations. Height growth rates changed dramatically during the growing season with two distinct peaks, one in the middle of May and the second at the end of July or beginning of August. However, the second peak in growth rates was only observed in certain populations that resumed faster growth after slowing down in June while the growth rates of the remaining populations were decreasing. A s a result of this contrasting behaviour, the largest differences among the populations in the average growth rates were observed at the end of July (GRJ). Variables related to plant size were highly correlated (SDW, R D W , T D W , H2, C A L ; 0.79 <r< 0.98); a weaker relationship was observed only between R D W and H2 (r = 0.67, P<0.05). G R J was well correlated with S D W (r = 0.73) but not with R D W (r = 0.47, P>0.1). P A I and S T E M S were strongly related to each other (r = 0.97) and displayed no significant 3 The results presented pertain to the 1997 experiment. The 1996 experiment is referred to but the results are not shown in detail. Al l provenances but one from the 1996 experiment were included in the 1997 experiment. 52 correlation with any other morphological variable. There was no statistically significant relationship between H I and H2 (P > 0.1). Table 7. Overall mean, standard deviation (SD), means' range and relative range (RR = range 100/mean), intraclass correlation coefficients (tp) of variables related to morphology. Overall mean and standard variation were calculated using all individual observations. Range and relative range were calculated for population means. Variable Unit Overall mean SD Range R R (%) tp H I cm 12.9 4.8 6 . 9 - 21.9 116.4 0.61 H2 cm 79.0 22.3 3 8 . 6 - 107.1 86.8 0.45 C A L mm 10.3 1.6 7 . 7 - 11.7 38.8 0.39 S D W g 17.6 7.6 5 . 1 - 29.2 137.5 0.52 R D W g 19.4 6.7 7 . 3 - 27.1 102.1 0.39 T D W g 37.1 13.2 1 2 . 4 - 55.4 115.9 0.50 R W R 0.53 0.07 0 . 4 1 - 0.59 32.1 0.26 S T E M S 1.62 1.02 1.04- 3.17 131.5 0.32 G R J cm/day 0.42 0.34 0.007-0.918 216.9 0.35 P A I 0.53 0.83 0.027-2.013 375.5 0.33 Patterns of variations Results of the multiple linear regression analysis between location and morphological variables are included in Appendix 9. H I was not correlated significantly with any of the location variables while plant architecture variables ( S T E M S and PAI) were correlated very weakly. Modest R 2 values were found for: H2 (0.45 with L A T and E L E V ) , C A L (0.39 with L A T and E L E V ) and R D W (0.41 with C O A S T , L O N G and E L E V ) . A good trend was detected for S D W (R 2 = 0.64) which was changing according to equation: S D W = 36.83 - 0.0214(LAT)-0.0079(ELEV) +0.0073(LONG). Most of the explained variation in S D W was due to latitude (partial r 2 = 0.38, P O . 0 1 ) and elevation (partial r 2 = 0.18, P<0.01). 2 2 The best clinal geographic trend was found for G R J (R z = 0.70) and for R W R (R z = 0.67). G R J decreased with latitude (partial r 2 = 0.57, P O . 0 1 ) and elevation (partial r 2 = 0.07, P O . 0 5 ) and increased with longitude (partial r 2 = 0.06, P<0.05). Northern populations invested relatively more carbon into roots compared to southern seed sources (partial r = 0.60, Figure 18). 53 R W R also increased, though weakly, with elevation (partial r = 0.07, P<0.05). It should be noted, however, that higher R W R in northern populations may have been caused by the continuous growth of roots in fall after their shoot growth was arrested. Correspondingly to the multiple linear regression analysis, canonical correlation analysis based on population means (Appendix 10) revealed strong latitudinal differentiation in variables related to plant size. However, a substantial coast-interior pattern of variation was also evident and was more clearly demonstrated than in the regression analysis. Out of four constructed canonical correlations, the first two were statistically significant (P<0.01) and accounted for 43% of the total variance of morphological variables. Both, L A T and C O A S T loaded highly (r = 0.84 and 0.7, respectively) on the first geographic canonical variable ( L O C A T 1 ) which accounted for almost all of the explained variance of morphological variables (88%). S D W , T D W , G R J and R W R were well correlated with L O C A T 1 (r = -0.79, -0.72, -0.82, 0.80, respectively). No variable loaded highly on the second and third geographic canonical variable. L O C A T 1 approximates temperature regimes and therefore length of the growing season as it increases with L A T and C O A S T and, to a lesser degree, with E L E V . 54 120 i • = : 1 01-May 01-Jun 01-Jul 01-Aug 01-Sep D A T E Figure 16. Changes in mean population height from 5 A p r i l to 25 September, 1997. 55 Figure 17. Changes in mean population height growth rates (cm/day) during the growing season. 56 Figure 18. Mean population root weight ratio ( R W R ) plotted against the latitude of the seed collection sites. Dashed line is based on regression with all populations excluding Californian population. 57 4.5 Relationship between the data sets Speed of germination and height after one growing season were not significantly correlated. Moreover, no strong linear correlation was found between any of the germination parameters and other variables. However, a significant relationship (r = 0.79) and good fit was found between log(GC) and S T E M S (Figure 19). Populations with lower germination capacity tend to have more bushy form, a possible hormonal effect. A statistically significant (PO.01) negative correlation was found between variables related to photosynthesis and mean population dry weights (e.g. r = - 0.75 between A and T D W ) . A strong negative relationship between the biomass and transpiration rate and gs was also detected (e.g. r = -0.75 between E and T D W ) . It should be noted, however, that this might have been a pot effect since large plants possibly had less water per unit biomass than smaller plants had as all seedlings were grown in the same size containers. A positive correlation was found for W U E and plant size (e.g. r = 0.79 with R D W , 0.72 with T D W ) . A statistically insignificant correlation was found between A and G R J (r = -0.46, P>0.1). Dark respiration rate was not significantly related to any other variable i n the 1997 experiment. In 1996 (ten coastal populations only), there was a good correlation between R D W and R (r = 0.93). Biomass accumulation was well correlated with G R J (e.g. r = 0.73 with SDW). Populations with lower height growth rates in the second part of summer had higher R W R (r = -0.80). Populations with higher G R J have a tendency to develop frost hardiness later and reach lower maximum frost hardiness, but the relationship was not very strong (r = 0.57, P<0.1 with FII on 11 November; r = -0.57, P O . l with January LT50). A strong correlation was found for time of bud flushing and time of frost hardiness development with r = 0.87 between bud flushing date (DAYS5o%budbreak) and FII evaluated at 58 -25°C on 11 November (Figure 20). Plants that developed frost hardiness later (i.e. had higher FII in fall) had a tendency towards late bud break in spring. Two dimensional mosaic plots (Figure 20) show close resemblance between the spatial patterns of variation of bud flushing date and fall frost hardiness. 4 (H 1 1 - i ' 1 0 20 40 60 80 100 GC(%) Figure 19. Mean population germination capacity (GC) and the average number of major stems (including the main stem) per plant (STEMS) . 60 3600 3500 3400 3300 1 3200 3100 3000 2900 DAYS • 30 25 20 • 15 i 10 ,^  ^  ^  ^ ^ ^ ^ <P ^  ^ # 1& i& ^ Longitude (min) (A) Longitude (min) (B) Fl! 84 72 60 48 36 24 12 Figure 20. Frost hardiness development and bud break. FII = frost injury index (%) evaluated on 11 November at - 25°C. DAYS>o<7rbudbrcak=number of days from 15 March for 50% of population plants to start bud break. (A) map of bud break timing in British Columbia, D A Y S = DAYS5o<7rbudbreak; (B) map of November frost hardiness in British Columbia. Note: to convert minutes of latitude and longitude to degrees divide by 60. 61 5.0 D I S C U S S I O N 5.1 Germination The very large differences observed in germination parameters resulted from genetic differentiation of populations and/or genetic-environmental interactions. Both, genetic and environmental factors can affect percentage of viable seeds in the sample, germinability of the viable seeds and seed vigor. The following discussion of environmental factors is restricted to the period of seed development only. Possible effects of seed handling and storage are noteworthy, but should be minimal since the collection procedure was uniform. M u c h larger differences in seed characteristics can be expected from site to site due to site-specific environmental conditions prior to seed collection. There seems to be a pattern of large variability in seed viability as well as frequent occurrence of low seed viability for Alnus spp. based on literature reports for alder species other than Sitka alder. Percent of viable red alder seeds, varied from 7.3% to over 74 % in one germination test (Elliott and Taylor, 1981) and from 59 to 87% in another test (Radwan and De Bel l , 1981). Percent of empty seeds for gray alder (Alnus incana spp. rugosa (Du Roi) J. Clausen), and black alder (Alnus glutinosa (L.) Gaertn.) was above 40% (Schalin, 1968). Large population variation in the seed viability (0-80%) for black alder was also reported by McVean (1955) and was due almost wholly to the failure of embryo formation (empty seeds). A sampled population of green alder (Alnus crispa (Ait.) Pursh.) in Alaska had 40%> viable seeds (Zasada et al., 1983). These studies indicate that the large differences in G C found in Sitka alder are typical of alder species. However, it is not clear what caused the differences: empty seeds, defective embryos, dormancy or population variable requirements with respect to germination conditions. A l l these factors can be influenced by the genetic makeup of the population as well as the local 62 environmental conditions during seed development. Percent of viable seeds may be affected by an early selection during seed development. Embryo abortion can be caused by high genetic loads from which many outcrossers suffer (Weins et al., 1987). Genetic load is defined as "relative lowering o f the mean fitness o f a population compared with the fitness of the best possible genotype" and is a consequence of the presence of mutated harmful recessive alleles (Charlesworth, 1989). Genetic load is readily expressed at various levels of inbreeding leading to the development of defective seeds (Park and Fowler, 1984; Sorensen and Cress, 1994). Since the rate of mutation to lethal form is not uniform for all locations, the genetic load carried in the population w i l l also vary with site (Bishir and Namkoong, 1987). In the case of Sitka alder, both site effects on mutation rate and inbreeding can play a significant role. This supposition follows from the species' very wide and patchy distribution range, its pioneer character that increases the chance of founder effect, and from population isolation due to the high elevation growth. In addition to genetic factors, environmental conditions can strongly affect seed viability. Stress from low/high temperatures, low/high moisture, limited resources and competition can all have an impact on seed development (Owens, 1991). Even very short exposure to adverse environmental conditions can hinder proper seed development by affecting the vigor, viability and quantity o f pollen, and can lead to the formation o f parthenocarpic seeds. Differences among the populations in germination completeness may have resulted not only from the presence of defective seeds but also from the presence of viable seeds that did not germinate. This situation could occur i f the seeds were dormant and the conditions prior to the germination test were not appropriate to break the dormancy or, less likely, the seeds were in a quiescent state imposed by inappropriate conditions during germination. Again, both genetic and environmental factors can play a role. 63 Various environmental conditions during seed development can affect the germinability of the viable seeds. Heide et al. (1976) reported that for red beets (Beta vulgaris L . ) , low temperature during seed development increased the percentage of empty seeds and decreased the germination capacity of viable seeds. Nooden et al. (1985) found that lack of mineral nutrients and cytokinin, a condition that can be caused by stress such as drought, results in a thicker, less permeable seed coat in soybean (Glycine max [L.] Merri l l) . On the other hand, Smith (1976) did not find effects of parent plant nutrition on germination capacity in lettuce (Lactuca sativa L.) seeds but only on the total seed yield. Environmental preconditioning effects on seed germination can be more subtle than rather obvious moisture or nutrient effects. Germination of Piper auritum, a rain forest pioneer tree, was dependent upon light conditions during seed maturation (Orozco et al., 1993). High or very low red:far red light ratio (R:FR) during seed development caused partial germination inhibition in seeds germinated at different R : F R than those during seed maturation. This phytochrome-mediated sensitivity o f seeds to light conditions was also shown in red alder, where germination is inhibited by far red light and stimulated by red light (Bormann, 1983; Haeussler and Tappeiner, 1993). Tanaka et al. (1991) suggest that the differences in germination parameters observed for red alder populations may have been caused by local environmental conditions. Population environmental conditions have also been suggested as a possible partial cause for differences in germination parameters for western hemlock (Campbell and Ritland, 1982), Pacific silver fir (Davidson et al., 1996) and paper birch (Bevington, 1986). In addition to environmental factors, stand maturity may also play a role since seed germination can be affected by the age of the mother plants (Gutterman, 1992). Large differences in germination parameters among Sitka alder populations indicate that the germination completeness as well as germination speed in this species may be adaptive. Genetic control of germination traits has been shown for Douglas-fir (El-Kassaby et al., 1992), 64 Pacific silver fir (Davidson et al., 1996), Sitka spruce (Chaisurisri et al., 1992), palebark Heldreich pine (Giannini and Bellari, 1995), aspen (Gallo, 1985) and yellow poplar (Barnett and Farmer, 1978). Between population differences in germination capacity may result from the variable levels of dormancy. Dormancy prevents seed germination in situations that could lower the survival rate of the seedlings; for example, where short periods of warm temperatures are followed by cold weather. Such unfavourable circumstances are often site-specific. In addition, long-term seed dormancy can improve reproductive success, particularly in less stable environments. Since Sitka alder often gets established after disturbance, seed banks consisting of seeds at variable dormancy would help this plant to colonize new areas. In addition, seed production may be very low or none in some years, increasing the importance of seed banks. There is no information with regards to Sitka alder seed dormancy. In this experiment, some populations had very high GC indicating that they were not dormant and stratification was not needed, at least for these seed sources. It has been suggested that red alder seeds are quiescent rather than physiologically dormant; i.e. would germinate in favourable conditions (Kenady, 1977; Radwan and De Bell, 1981). However, Elliott and Taylor (1981) showed that while physiological dormancy in red alder seed is not common, it does exist in some populations. It is probable that the same situation can occur in Sitka alder, especially that it grows in less predictable, high elevation environments and has wider distribution range than red alder. Germination speed of viable, non-dormant seeds, measured by R'50, may also be an adaptive trait like germination completeness. The adaptive significance of higher/lower speed of germination could be to improve plant survival by minimizing the exposure to stress (e.g. drought, low temperature) or by taking advantage of short favourable growing conditions. For example, Weber and Sorensen (1992) found that the germination speed of ponderosa pine is highest in populations from regions of short, drought-limited growing seasons. In this study, 65 northern seed sources of Sitka alder germinated faster than southern. This is consistent with literature reports for some species: e.g. paper birch (Bevington, 1986) and western hemlock (Campbell and Ritland, 1982). Faster germination in the north may be an adaptive trait since the benefits of extending the length of the growing season by several days may outweigh the risk of exposure to low temperatures. The opposite may be true in the south where the growing season is longer and delayed germination by several days w i l l not affect the seasonal growth rate but may improve survival rate of the seedlings. Environmental and genetic effects on germination parameters, confounded in this experiment, produced very large differences among the populations since over 90% of the total variance was due to population effects. Although it is not possible to determine the contribution of each factor to the total variation, some hypotheses can still be offered with respect to the pattern of variations. Variations in germination completeness of the studied Sitka alder populations seem to be random since no substantial geographic trend was found for G C . For example, Cypress Park and Hemlock Valley, populations o f virtually identical elevation and close proximity, had strikingly different G C (15% and 76%, respectively). Therefore, i f such variations in G C are adaptive or environmental in nature, they must be due to some small scale environmental conditions. On the other hand, the differences may also have been caused by random genetic drift. Speed and germination completeness responded differently to' environmental pressures. There was no correlation between G C and R50' implying that the long or short term environmental conditions that possibly affected quantity of the germinated seeds did not affect the vitality of these seeds. Also , the probable environmental preconditioning effects on seed vigor were not lasting since there was no correlation between seedlings height after one growing season and any of the germination parameters. Similar results were found for aspen (Gallo, 66 1985). If, on the other hand, low G C in some populations was due to inbreeding leading to the expression of harmful recessive alleles during embryo development, the inbreeding did not affect the vigor of the seeds that germinated and the subsequent growth of the seedlings. 5.2 (a) Frost Hardiness Large differences among the populations in fall and winter frost hardiness indicate strong genetic control of this attribute. The differences were particularly large during the time of frost hardiness development where over 65% of the total variance was due to the population effect. Bigger differences in fall frost hardiness compared to the differences in maximum frost resistance are often observed (Sakai and Larcher, 1987). For example, Deans and Harvey (1996) found the largest interpopulation differences in sessile oak (Quercus petraea (Matt.) Lielbl.) frost hardiness in fall and spring and the smallest ones in midwinter. In contrast to that study, no significant differences were found for Sitka alder cold resistance in spring. The range of means for the last test in March was very close to the range observed for several fall and winter tests that produced significant differences. However, much higher variations within populations occurred in March and, as a result, the populations were not significantly different. Similar results were found in red alder (Cannell et al., 1987) where northern populations started to develop cold resistance earlier in the fall, but all populations dehardened at about the same time in March. The geographic trend in Sitka alder frost hardiness is rather unusual. A s a group, interior populations tend to be somewhat more frost hardy than coastal populations; however, populations from the interior wet belt of the western side of Rocky Mountains were frost hardy only to the same levels as coastal plants. The latitudinal trend in frost hardiness, often reported for other species (Flint, 1972; Joyce, 1987), was distinct in Sitka alder but not as strong as might 67 be expected from the large differences in population latitudes (18 degrees). The trend was more evident in fall frost hardiness than in the maximum cold resistance. In contrast, Sakai (1983) reported a close relationship between midwinter low temperatures and plant adaptation to cold climates. It seems that out of the two main factors, temperature and photoperiod, known to affect development of frost resistance (Kramer and Kozlowski , 1979), temperature has a stronger effect on the onset of frost resistance development and its progress in Sitka alder. This hypothesis is based on deviations from the latitudinal trend and low levels of frost resistance developed in fall as well as in midwinter. The observed latitudinal trend in Sitka alder November frost hardiness implies that photoperiod may be involved in the induction of frost resistance. The deviations from the trend indicate that the response to photoperiod relates more to the local temperature regimes than to latitude or that some other factors that do not change regularly with latitude are important in frost hardiness development. For example, V N , the most frost hardy population, comes from the middle of the latitudinal range of the species in British Columbia. Southern seed sources, C R and P H were as frost hardy at the beginning of November as C H and D E , both located north of Prince George. In a study with black alder, red alder and green alder, Tremblay and Lalonde (1987) found that frost hardiness in these species was more affected by low temperature than by photoperiod. These alders developed similar levels of frost hardiness in low temperature treatment (5°C) regardless of photoperiod, although there was a small increase in cold resistance in plants from northern populations grown in warm temperatures (22/18°C) and exposed to short days (8 hours). Plant frost hardiness reflects adaptation to temperatures which in general decrease with 68 latitude but may also show significant deviations from that pattern depending on the number of factors such as distance from the coast, topography and precipitation. These deviations may result in the departures from the latitudinal pattern in Sitka alder frost hardiness. It also means that some southern seed sources could respond to the same photoperiod (that occurs earlier in the south for days longer than 12 hours) as northern populations i f the decrease in temperature in the south occurs earlier than in the north. Naturally, populations of the same latitude are expected to respond to different photoperiods i f they come from locations different with respect to altitude or distance from the coast. This reasoning can explain lack of a complete correlation of frost hardiness with latitude and therefore with photoperiod even i f such a relationship between night length and cold resistance induction may exist. However, there is still some indirect evidence that photoperiod may not be as important as temperature. One indication comes from an observation that species that develop an extremely high level of cold resistance in winter w i l l respond to short photoperiod by hardening even without the low temperature treatment (Sakai and Larcher, 1987). On the other hand, species that are not very frost hardy are often weakly affected by photoperiod (Williams et al., 1972; S i l im and Lavender, 1994). The levels of maximum hardiness attained by Sitka alder, as reported here, are not as great as might have been expected for a high elevation species, even though the population sample included northern and interior seed sources. The average LT50 for all populations in January was only -30 °C (ranging from -20 to -50 °C). However, it is also possible that these plants did not attain higher frost hardiness because of the relatively mi ld climate in Saanichton (although winter 96/97 was unusually cold with temperatures dropping to below -10 °C). I f that is the case, then the real frost hardiness was underestimated and the differences among the populations may be greater in lower temperature regime than were found in Saanichton. 69 Another argument for a lesser role of photoperiod is based on initial frost hardiness development by Sitka alder populations in the test conditions of south Vancouver Island. Assuming that frost hardiness developed first in populations from colder regions, three scenarios with respect to frost hardiness induction can be presented: (1) "cold" populations responded to higher temperature than "warm" populations, (2) both responded to the same temperature but "cold" populations developed frost hardiness faster, (3) populations responded to site-specific photoperiod. The first scenario seems to be the most likely: seed sources from colder regions develop frost hardiness in response to the change in higher temperatures than "warm" populations because of the risk of sudden drop in temperatures in early fall. For example, temperatures can drop to -12 to -15 °C in September in northern British Columbia (Climatic atlas, 1984). Figure 5 suggests that once the population-specific environmental cue is received, the populations develop frost hardiness at similar rates. For example, the slopes describing changes in FII of the most and the least frost hardy population are identical during rapid frost hardiness development. Slopes for many other populations are similar even though frost hardiness was likely to be estimated at a different developmental stage for each population. The role of photoperiod in frost hardiness development (third scenario) also seems to be small based on low levels of October frost hardiness in some populations: the five most northern interior populations would survive temperatures in the vicinity of only -12.7 °C, while coastal populations could endure -10 °C (data not presented). October frost hardiness of coastal populations would be enough to survive in their natural locations. However, northern populations were definitely not frost hardy enough to endure temperatures as low as -25 °C that can occur in October north of Prince George (Climatic atlas, 1984). That is despite the fact that these populations were exposed earlier in the season to the same photoperiods as they would be in their natural location since the experiment was located south of all seed collection sites but one. A s a 70 result, the "cold" populations should have developed frost hardiness much faster than "warm" populations because they should respond to the longer daylength than "warm" populations and that particular daylength occurs earlier in the south (prior to 21 September) compared to the northern locations. Accordingly, one would also expect larger differences among the populations in October than were found if photoperiod was strongly involved in cold resistance development. The same level of cold hardiness of populations from regions of high snowfall (coastal regions and interior wet belt), relatively low frost resistance and its weak geographic trend in midwinter all indicate that snow cover may influence adaptation of Sitka alder to winter temperatures. Snow is an excellent insulating material. For intermediate latitudes the temperature usually does not decrease below -5 °C beneath snow cover thicker than 20 cm (Sakai and Larcher, 1987). Important for this hypothesis is the small size of Sitka alder that can be as low as 3 m for mature plants (Haeussler et al., 1990). Also, its bushy growth form observed in natural stands may result from frost damage to the leader and main stems that were not protected from low temperatures by snow. 5.2 (b). Frost hardiness and bud break Strong correlation was observed between fall frost hardiness and date of bud break. Plants that develop frost resistance earlier also break their buds earlier in spring. It is not possible, based on this study, to determine whether frost hardiness and dormancy are interdependent or correlated without causation in Sitka alder. In general, the relationship between dormancy and cold hardiness is not clear since there are species that do not develop dormancy yet develop frost resistance (Silim and Lavender, 1994), or develop dormancy but not frost resistance (Kramer and Kozlowski, 1979). On the other hand, frost resistance and dormancy are often positively correlated in many species (e.g. Fuchigami et al., 1982; Sakai and Larcher, 1987; 71 Kuzina and Kalinina, 1993; Erstad, 1994). There may be several reasons for the differences among the populations in time of bud break: different time of dormancy onset, different chilling requirements to break bud dormancy or different heat requirements for bud flushing after dormancy is broken. Since Sitka alder populations that develop frost hardiness earlier also break their buds earlier, and since frost resistance and dormancy are often positively correlated it follows that populations that develop dormancy earlier break their buds earlier. This could mean that the duration o f dormancy was similar for all populations in this common garden experiment and the differences in bud break resulted from the differences in time of dormancy development. The relationship between time of growth cessation and time of growth resumption has been apparent in nurseries that use short-day treatment. It was observed that the short-day treatment applied to arrest height growth and induce frost resistance also induces early bud break (Kozlowski and Pallardy, 1997). On the other hand, the duration of dormancy may not be related to the date of dormancy onset but rather to chilling requirements needed to break dormancy. Different chilling requirements of populations could therefore affect the timing of bud break. For example, chilling requirements of sweet cherry cultivars (Prunus avium L.) were positively correlated with late date of bud opening (Seif and Gruppe, 1985). Since Sitka alder populations from colder regions generally started to grow earlier they would have to have lower chilling requirements than seed sources from warmer regions. Such adaptation has been shown for silver birch (Betula pendula Roth.) in which the chilling requirement decreased with increasing latitude of origin (Myking and Heide, 1995). However, the opposite has also been shown in other species (Nienstaedt, 1967; Kriebel and Wang, 1962). This apparent discrepancy may relate to the temperature regimes of sampled seed sources: plants from warm regions with fluctuating warm and cold temperatures would require longer exposure to cool non-freezing temperatures (i.e. a temperature of around 72 5°C most effectively satisfies chilling requirements (Bigras, 1996)) in comparison to plants from cold regions characterized by prolonged periods of below-zero temperatures. For example, coastal populations of Scotch pine and silver birch had higher chilling requirements than the interior populations of these species (Leinonen, 1996). On the other hand, chilling requirements of continental seed sources may increase with latitude to prevent bud flushing while the temperatures are still low. There are many reports in the literature of northern populations starting growth earlier than southern (warm) seed sources in common garden experiments (Morgenstern, 1969; Steiner, 1979; Brissette and Barnes, 1984), but there are also examples of the opposite trend (Ashby, 1992). Yet another reason for differences in time of bud break in Sitka alder populations may be related to different heat requirements rather than to variable dormancy. It has been suggested that the chilling requirement plays a role in preventing bud break in fall (Borchert, 1991) rather than preventing too early a flush in spring. After chilling requirements are met the plants become quiescent; that is, their "dormancy" (called "ecodormancy") is imposed by adverse environmental conditions rather than by physiology (Kozlowski and Pallardy, 1997b). In order to start to grow the plants must be exposed to warm temperatures for a certain period of time. Different populations may have different heat requirements and therefore start to grow at different dates in a common garden experiment. Based on the following: (1) frost hardiness usually correlates with dormancy, (2) there were no differences among Sitka alder populations in frost hardiness in March, and (3) some of the populations were already starting to break buds in early March, it can be concluded that all populations were not dormant at the beginning of March even though most of them started to grow several weeks later. Grey alder and black alder vegetative buds were also released quite early from dormancy (in February) in an experiment in Norway (Heide, 1993). Therefore, the likely cause of differences in timing of bud break in Sitka 73 alder is differential heat requirements needed by buds to start to grow. Plants from colder regions would have lower heat requirements in accordance with lower spring temperatures in their native habitats and, as a result, would flush earlier in a common garden experiment established in a warmer climate. On the other hand, it is also possible that some populations started to grow before others because they were released from dormancy earlier and were thus able to start accumulating their heat sum earlier. 5.3 (a). Gas Exchange Genetic variations in the instantaneous photosynthetic rate of Sitka alder are related relatively well to geography as over 50% of the total variance in A was due to the location variables. Close to 80% of the explained variation was ascribable to coast versus interior differentiation. The cluster analysis grouping of coastal (low A ) and interior (high A ) populations into two separate groups with respect to A suggests ecotypic variation in A over large geographic areas. This kind o f broad ecotype is commonly reported in the literature for forest trees for other characteristics (Morgenstern, 1996). Accepting variation in A as ecotypic may be reasonable from the practical point of view; e.g., for seed transfer. However, it may be misleading i f an adaptive interpretation of the variations is required. Ecotypic variation means that there is no gradual change in the trait over a gradient of environmental conditions. This scenario seems unlikely for the rate of photosynthesis. Rather, the changes in environmental selective pressures were not well represented by the variation in the location variables. For example, the change in environmental conditions with the distance from the coast is more likely to be non-linear rather than linear and characterized by small differences in conditions within a certain distance from the sea, a steep environmental gradient over the mountains and more uniform again across the interior plateau. Thorough sampling in the transition zone of the steep environmental gradient 74 may in fact show that the variations in A are clinal. Regression analysis revealed that A increased not only with the distance from the coast, but also with latitude and elevation. Although the effects of latitude and elevation were small compared to the effects of the distance from the coast, the overall pattern of variation indicates that the maximum rate of photosynthesis tends to be higher in populations from regions with shorter growing season. While there are not many reports of significant geographic trends in A, there are some that indicate that my results are not unique. Higher rates of photosynthesis for high altitude populations are reported for Scotch pine (Zelawski and Goral, 1966), knobcone pine (Pinus attenuata Lemm.) (Wright, 1971) and sugar maple (Acer saccharinum L.) (Ledig and Kurbobo, 1983). Northern seed sources of Norway spruce (Gross and Hettesheimer, 1983) and black spruce (Picea mariana (Mill.) B.S.P.) (Johnsen et al., 1996) had higher A than southern populations during certain periods of the growing season. A clinal trend in the rate of photosynthesis was found for Scotch pine in Europe (Luoma, 1997). Photosynthetic capacity increased with latitude up to 57° north and declined for populations over 60°. However, the genetic and environmental effects were confounded in this study as the trees were measured in natural stands rather than in a common garden. A weak but significant increase in A from southeast to northwest was reported for red alder (Dang et al., 1994). Dunlap et al. (1993) report that clones of black cottonwood from the maritime climate in Washington state had lower A than clones from the continental climate. Douglas-fir populations from the interior had 27% higher A than did those from the coast (Zhang et al., 1993). The results indicate that populations from colder regions may have adapted to the shorter growing season by evolving higher maximum rates of photosynthesis than are observed in the "warm" seed sources. It may seem that high A would be an advantage in any situation regardless of population location. However, it is also possible that leaves with higher rates of 75 photosynthesis have a shorter life-span. Different leaf longevities in Sitka alder populations, though not estimated, were noticed as coastal populations still had leaves late in fall when populations from colder areas shed all their leaves. Differences among the populations in time o f leaf shed were shown for other species such as eastern cottonwood by Y i n g and Bagley (1976) and northern red oak (Quercus rubra L. ) by Kriebel et al. (1976). Although this apparent trade-off between leaf retention and rate of photosynthesis has not previously been reported to exist within species, it has been shown to occur between species. In a common garden experiment, Gower et al. (1993) grew five species of different leaf longevities: northern red oak, European larch (Larix decidua M i l l . ) , eastern white pine (Pinus strobus L.) , red pine (Pinus resinosa Ait . ) and Norway spruce. They found a strong inverse correlation between leaf life-span and maximum net photosynthesis per unit mass; no correlation was found when photosynthesis was expressed per unit area. A similar inverse relationship between leaf longevity and photosynthesis was found for a number of conifers as well as broad-leaved species grown in different sites (Reich et al., 1992). Reich et al. (1995) indicated that such an inverse relationship holds true for angiosperms even i f photosynthetic rate is calculated per unit area. Also , leaves that are grown in lower light and therefore have lower A , last longer than those grown in full light (Reich et al. 1995). It is not known i f the inverse relationship between leaf longevity and photosynthesis, expressed as the maximum rate of assimilation, would hold true i f photosynthesis was integrated for the entire growing season. It is likely that leaves that achieved lower rates of photosynthesis and last longer would have higher carbon gain per growing season. Johnsen et al. (1996) reported that the most northerly population of black spruce they examined had the highest rate of A during mid-summer and displayed the earliest decline in A in fall. 76 5.3 (b). Gas Exchange and Biomass Accumulation A negative correlation between the rate of photosynthesis expressed per unit leaf area and biomass accumulation was found for Sitka alder but was not completely unexpected. Negative or non-significant relationships between rate of photosynthesis and plant size parameters are often reported in the literature (e.g. Sorensen and Ferrell, 1973; Mebrahtu and Hanover, 1991; Zhang et al., 1993). Rate of photosynthesis is only one of several factors that can affect biomass accumulation. Other factors include maintenance and growth respiration rates, total leaf area per plant, leaf longevity, length of the growing season, integrated photosynthesis over the entire growing season and growth rate pattern during the growing season. In addition, carbon allocation to fine roots and their turnover, carbon loss due to exudates and mycorrhizae can play a role since these processes are rarely taken into account while reporting plant size parameters. In the 1996 experiment, leaf area and biomass were positively but not strongly correlated. Leaf dark respiration rates measured for all populations in 1997 were not significantly correlated with any of the biomass parameters. This observation implies that either respiration rate was not a significant factor for biomass accumulation, or that respiration was measured on the wrong tissues or at the wrong time with respect to biomass accumulation (e.g. perhaps shoot or root respiration would give different results), or that short term measurements of respiration are not sufficient. Variable success in terms of attempts to relate growth and respiration rate is reported in the literature (e.g. Anekonda et al., 1994). It is highly probable that the differences among the populations in biomass accumulation mainly resulted from the different length of the growing season. This hypothesis is based on the observed large differences in height growth rates in late summer. Populations from colder regions stopped growing in height earlier than southern populations which means that they set their buds earlier and stopped adding new leaves. As a result, the effects of the longer growing 77 season are particularly pronounced, since not only is carbon assimilated for a longer period of time but also smaller proportion of aged leaves, less efficient in carbon assimilation, is present in late summer and fall in populations with longer growing season. Kozlowski and Pallardy (1997a) reported that the rate o f photosynthesis decreases with leaf age. Large differences in height growth among populations of loblolly pine (Boltz et al., 1986) and jack pine (Pinus banksiana Lamb.) (Logan, 1971) were attributed to late season growth and high late season photo synthetic rate. Although an inverse relationship was found between A and dry weight, a good correlation was found between W U E and biomass. Populations with higher W U E tend to accumulate more biomass. It is apparent that plants with higher A also had a disproportionately high gs, resulting in low W U E for populations with high A . Strong positive correlation between intrinsic water use efficiency and productivity and no correlation with A was found by Zhang et al. (1993) for Douglas-fir populations. They speculated that the intrinsic W U E may be correlated to some variable that is more important in determining productivity than the absolute photosynthetic rate. Positive correlation between productivity and W U E (estimated using carbon isotopes) was also found by Holowachuk (1993) for lodgepole pine and by Sun et al. (1996) for white spruce (Picea glauca (Moench) Voss). 5.3 (c). Water relations Highly significant differences among the populations in stomatal conductance and intrinsic W U E serve as evidence of differentiation among Sitka alder populations with respect to water use. However, the adaptive significance of the differentiation is less clear. The observed trends in both gs and W U E , though significant, were not particularly strong. Out o f the total variance, 34% for gs and 29% for W U E was explained by the population location variables. This 78 was not entirely unexpected because of the complicated geography of British Columbia. Nevertheless, some differences between coastal populations and populations from the more drought prone interior were found, particularly for gs, but the nature of the differences was counterintuitive. Xeric populations did not conserve water under the normal moisture regime provided. Populations from the interior had higher gs and therefore lower W U E than coastal populations, although the relationship for W U E was much less strong. Similar results have been reported for other species. Seedlings of loblolly pine (Bongarten and Teskey, 1986), green ash (Fraxinus pennsylvanica Marsh.) (Abrams et al., 1990), Aleppo pine (Pinus halepensis M i l l . ) (Tognetti et al., 1997), Douglas-fir (Zhang et al., 1993) and black cottonwood (Dunlap et al., 1993) from xeric locations had higher stomatal conductance than seedlings from mesic environments in common garden experiments. Interior plants maintained greater gs than coastal plants at the same temperatures (Figure 13) and the same R H of air entering the cuvette. Higher gs and therefore higher E lowered V P D in the cuvette for these plants thus creating a positive feedback since lower V P D increases gs. Nonetheless, the fact is that interior populations responded differently to the same initial conditions in the cuvette implying that they can have higher gs than coastal populations at the same V P D . This relationship was found for the lower range of V P D (Figure 14). It does not mean, however, that the stomata of plants from both groups respond differently to the same environmental conditions. The higher or lower gs may be due to the differences in morphological characteristics related to stomata (number of stomata per unit leaf area, stomatal distribution, size of the stomatal opening) or characteristics related to leaf surface (colour, ability to reflect light, presence of hair). There may be a number o f reasons why Sitka alder from xeric regions did not show adaptation to conserve water. A n obvious possibility is that plants with higher gs had lower total 79 leaf area per plant so in fact they could loose less water per gram of total biomass. However, based on the 1996 experiment, there was no significant correlation between leaf area and gs (r = -0.36). Another potential explanation relates to the fact that plants in this experiment were well watered and it is possible that a greater ability to conserve water would be evident had the plants been subject to drought. On the other hand, the opposite reaction is also possible, as Tognetti et al. (1997) showed that Aleppo pine seedlings from more xeric regions had actually greater gs than mesic populations when both were subject to water stress. They proposed that xeric populations are more resistant to desiccation so they can afford to loose more water. High gs would also be advantageous for thermal regulation on hot summer days (if soil moisture content is high) as suggested by Dunlap et al. (1993) for black cottonwood. Kolb and Robberecht (1996) found that seedlings of ponderosa pine that survived drought and high temperatures had significantly higher stomatal conductance than seedlings that did not survive. They also reported that seedlings subject to drought stress and root competition at the same time were injured more than those protected from competition, and those that survived competition had the same high gs as those that survived drought in the absence of competition. These results are consistent with the hypothesis proposed by Holowachuk (1993) that by not conserving water under drought conditions, plants with low WUE (high gs) may gain a competitive advantage by limiting water availability for other plants. 5.4 Quantitative attributes All studied quantitative traits related to carbon accumulation and allocation in Sitka alder were found to be under strong genetic control. Large population differences in various biomass parameters are commonly reported in the literature for many species (Morgenstern, 1996). The largest differences among Sitka alder populations were found for HI since over 60 % of the total 80 variance was due to the population effect. Smaller differences were detected for H2 (45 %) indicating that perhaps there were some maternal effects confounded with genetic effects thus causing larger differences in height during the first growing season. The adaptive significance of the variations in Sitka alder biomass parameters evaluated at the population level can be suggested for those traits that showed some geographic trend. Among the variables related to plant structure and carbon allocation pattern (RWR, S T E M S , PAI) only R W R showed a significant geographic trend, while S T E M S and P A I were not related to any of the location variables. Wi th the exception of several coastal populations, the form of the majority of plants was tree-like (monopodial) rather than bush-like (polypodial). In natural stands, however, the habit of Sitka alder is usually bush-like (Binkley, 1981). Perhaps the growth form changes with time and these plants w i l l become more bushy after several years or the bushy habit form results from damage to the main stem in natural stands (e.g. from frost, browsing or insects). In contrast to S T E M S and P A I , R W R showed a distinct geographic trend by increasing strongly with latitude and less strongly with elevation. Greater allocation of carbon to roots in plants from colder regions is often observed (Schultz and Gatherum, 1971; Cannell and Willet, 1976; M a , 1989; Stahl and Persson, 1992). Plants may invest more carbon into roots in conditions of low soil temperatures to compensate for lower nutrient (Ruel et al., 1996) and water uptake (Kramer and Kozlowski , 1979), in response to lower soil fertility and/or moisture availability (Keyes and Grier, 1981). Sitka alder "cold" populations may also have, higher R D W / S D W to increase their vegetative reproduction. Vegetative reproduction may be particularly important in cold regions where sexual reproductive success may be low in some years due to unfavourable environmental conditions and where above-ground parts may get injured or even ki l led by frost. The ability o f Sitka alder to sprout from roots was observed in this experiment and is also reported in the literature (Haeussler et al., 1990). 81 In general, the geographic pattern for variables describing plant size and biomass accumulation rate indicates that smaller plants come from colder regions. The trend was shown particularly strongly by the canonical correlation analysis for S D W , T D W and G R J which decreased with both latitude and distance from the coast. A similar trend, but less strong, was found for H2 and R D W . Smaller size of plants from colder populations is often reported in the literature (Brissette and Barnes, 1984; Rehfeldt, 1987; M a , 1989; Y i n g , 1991) and may be related to differences in growth pattern through the growing season. Average daily height growth rates were changing differently during the growing season for each population, although there was a distinct depression of growth rates in June for every population except one. Since June 1997 was particularly cold and since virtually all populations decreased their growth rate at the same time it follows that the growth depression might have been an environmental effect. Another explanation for the growth rate pattern might be periodic shoot growth with an intervening period of rest (Borchert, 1991). Close examination of the stems did not reveal, however, the presence of more than one set of bud scale scars per year, suggesting that recurrent flushing did not occur. After the growth depression in June, some populations resumed rapid growth while others continued slowing down. A s a result, at this time there were the largest differences in growth rate and a strong geographic trend in this trait. Higher growth rates were found in populations from warmer regions indicating a longer growing season. Large differences found among the populations in R D W and R W R indicate the possibility to select seed sources most suitable for slope stabilization and road deactivation. In the case of R D W , the relatively weak geographic trend found for this trait means that it is possible to find populations with high R D W within any given geographic region. For Vancouver Island and the south coastal region of the mainland these include G R , R O B , SP and O W . Since R D W and R W R were not significantly correlated, a different seed source must be selected i f 82 small plants with relatively large roots are required rather than plants with large R D W . Populations with large R D W w i l l also have large above-ground biomass which may have undesirable effects on commercial species grown in the same place. Because there was only a weak negative correlation between fall frost hardiness and biomass (r = 0.56 with S D W and 0.46 with R D W ) and no correlation with midwinter hardiness, it should also be possible to find populations with superior growth and high frost hardiness i f such a trait combination is required. A s indicated in the discussion of the gas exchange parameters, productivity of Sitka alder is probably largely related to the population-specific length of the growing season. This conclusion is based on the positive correlation between height growth rates in late summer and above ground biomass accumulation. The relationship between whole plant productivity and length of the growing season may be stronger than shown since height increments of the main stem are not an ideal measure of growth activities of the whole plant. After shoot elongation ceases, the shoot w i l l still increase its mass due to cambial growth, which often continues after bud set (Kramer and Kozlowski , 1979), and due to accumulation of reserve carbohydrates in parenchyma tissues (Chomba et a l , 1993). It is possible that the biomass accumulation after the height growth ceases wi l l be relatively large in plants that stopped growing in height later since they may have larger leaf area per plant. In addition, height growth rate does not reflect growth rate of roots since they may continue to grow long after bud set (Kramer and Kozlowski , 1979) and respond to different environmental cues. Also , the rate of growth of the main stem may be different from that of the branches since lateral shoots sometimes show a different pattern of seasonal growth (Kramer and Kozlowski , 1979). These two factors can further distort the relationship between growing season and plant productivity i f the length of the growing season is estimated using height growth rates of the main stem. 83 6.0 C O N C L U S I O N S A N D S U M M A R Y Large genetic differences among Sitka alder populations were found in germination parameters, fall and winter frost hardiness, bud break, characteristics related to gas exchange (photosynthesis, dark respiration, stomatal conductance, transpiration, carboxylation efficiency, and intrinsic water use efficiency) and characteristics related to biomass allocation (height, caliper, shoot dry weight, root dry weight, height growth rate and number and size of major stems per plant). N o differences were detected for spring frost hardiness and stomatal sensitivity to vapour pressure deficit. The populations differed significantly in fall and winter frost hardiness, though the differences were greater at the time of frost hardiness development. A l l plants that were subject to destructive sampling had root nodules formed by symbiosis with Frankia. The symbiosis could potentially distort the examined genetic variations. However, since the plants were fertilized with nitrogen and since the presence of nodules is not a sufficient evidence of nitrogen fixation, it was assumed that the nodules were not a significant factor in the study. Between 13-31% and 26-61%) of the total variance (sum of between and within population variances) was due to population effects for gas exchange and biomass related variables, respectively. Repeatability values for frost hardiness (0.89) and gas exchange variables (0.52 - 0.84) indicate that heritability of these traits may be large. Based on univariate (multiple linear regression) and multivariate methods (canonical correlation and cluster analysis), geographic patterns of variation were found for most traits. Variation related to geography was particularly strong for fall frost hardiness, shoot dry weight, growth rate in July, ratio of root dry weight to total dry weight and net photosynthesis rate. Most of the variation was related to latitude and distance from the coast. Geographic trends helped to identify the adaptive nature of some traits. Seedlings 84 originating from northern seed sources were in general more frost hardy, smaller and allocated more carbon to roots compared to shoots. Plants from the interior wet belt and the coast were less frost hardy than plants from the dry interior, suggesting that snow cover plays a role in Sitka alder adaptation to low temperatures. Photosynthetic capacity increased with distance from the coast and, to a lesser degree, with elevation and latitude, implying that an inverse relationship exists between the length of the growing season and the rate of photosynthesis. Interior populations had higher A due to higher carboxylation efficiency and stomatal conductance (gs). As a result of higher gs, plants from more xeric interior regions had a lower water use efficiency. Productivity was negatively correlated with A and probably depends more on the length of the growing season. Speed of germination did not affect the plant size after one growing season and was independent from germination completeness. A strong positive correlation was found between timing of frost hardiness development and bud break. Large differences among the populations and good geographic trends for some traits indicate that Sitka alder has enough variation within populations for natural selection to occur despite the species population dynamics characterized by rapid expansion and contraction that can erode genetic resources. It also indicates that the selection process is rapid in this high elevation species since Sitka alder is an early successional, short-lived pioneer plant. On the other hand, large genetic differences not related to geography (found for several traits) suggest that this species may be subject to random genetic drift or that the location variables did not reflect the selective forces. Differences among neighboring populations indicate the possibility of significant barriers to gene flow in addition to already mentioned genetic drift and small scale site heterogeneity. The observed patterns of large genetic variation in Sitka alder populations represent a challenge for gene conservation efforts in high elevation ecosystems. 85 7.0 F U T U R E D I R E C T I O N S This study left some questions unanswered and raised several issues worth investigating in the future. These include issues pertaining to Sitka alder as well as to population testing in common garden experiments. Addressing questions related to Sitka alder specifically is important not only in terms of understanding this species as an entity unto itself but also as a component of high elevation ecosystems. Large variations among the populations were found for most studied traits. The study was less successful, in many cases, in relating the inter-population variations to geographic variables (latitude, longitude, elevation and distance from the coast). This implies that additional geoclimatic data (such as temperature, precipitation, soil properties, wind patterns, exposure to sun) of seed collection sites may be needed to explain the differences among the populations, or that there is no geographic pattern of variation. For example, information on the depth and duration of snow cover could strengthen the proposed hypothesis that snow plays a significant role in Sitka alder adaptation to low temperatures. Additional experiments are needed to study extremely high variations in population G C and to test the hypothesis of an inverse relationship between A and leaf longevity expressed at the population level. Finding the basis for large variations in G C , a trait that Sitka alder shares with many other alder species, is important for utilization and conservation of this plant. Negative correlation between A and the length of the growing season may not be exclusive to Sitka alder and it could have substantial implications for breeding strategies. In addition it could explain the lack of relationship between A and biomass accumulation often found in other species. Symbiosis with nitrogen-fixing bacteria plays a crucial role in Sitka alder establishment 86 and survival in natural conditions. Thus, it may be worth looking at this plant adaptive attributes in experiments that include symbiosis with Frankia as a factor. Two possible artifacts of the common garden experiments, one with respect to the level of frost hardiness and second to R W R were noted. A relatively low level of Sitka alder frost resistance may be due to a higher temperature regime in the common garden experiment compared to the temperatures in natural locations. Higher R W R found in northern populations may have been caused by the continuous growth of roots in fall after their shoot growth was arrested in response to environmental signals. Both frost hardiness and R W R are consequential from the utilization point of view, so these potential artifacts should be scrutinized. 87 R E F E R E N C E S : Abrams, M . D . , Kubiske, M . E . and Steiner, K . C . 1990. Drought adaptations and responses in five genotypes of Fraxinuspennsylvanica Marsh.: photosynthesis, water relations and leaf morphology. Tree Physiol. 6: 305-315. Ager, A . A . , Heilman, P.E. , and Stettler, R.F. 1993. Genetic variation in red alder (Alnus rubra) in relation to native climate and geography. Can. J. For. Res. 23: 1930-1939. Anekonda, T.S., Criddle, R.S. , Libby, W.J . , Breidenbach, R . W . and Hansen, L . D . 1994. Respiration rates predict differences in growth of coast redwood. Plant Cel l Environ. 17: 197-203. Ashby, W . C . , Bresnan, D.F. , Roth, P .L . , Preece, J .E. and Huetteman, C A . 1992. Nursery establishment, phenology and growth of silver maple related to provenance. Biomass Bioenergy 3: 1-7. Barnett, P .E . and Farmer Jr., R . E . 1978. Altitudinal variation in germination characteristics of yellow poplar in the southern Appalachians. Silvae Genet. 27: 101-104. Bassman, J .H. , and Zwier, J .C. 1991. Gas exchange and leaf area characteristics of Populus trichocarpa, Populus deltoides, and Populus trichocarpa x deltoides clones. Tree Physiol. 8: 145-159. Becker, W . A . 1984. Manual of Quantitative Genetics. Academic Enterprises, Pullman, Washington. 190 pp. Bennett, K . J . , and Rook, D . A . 1978. Stomatal and mesophyll resistance in two clones of Pinus radiata D . Don known to differ in transpiration and survival rate. Aust. J. Plant Physiol. 5:231-238. Bevington, J . 1986. Geographic differences in the seed germination o f paper birch (Betula papyrifera). Amer. J. Bot. 73: 564-573. Bigras, F.J . 1996. Conifer Bud Dormancy and Stress Resistance. In Plant Dormancy: Physiology, Biochemistry and Molecular Biology. Edited by Lang, G . A . C A B International, Wallingford, U K . 386 pp. Binkley, D . 1982. Nitrogen fixation and net primary production in a young Sitka alder stand. Can. J. Bot. 60:281-284. Binkley, D . 1984. Douglas-fir stem growth per unit of leaf area increased by interplanted Sitka alder and red alder. For. Sci. 30: 259-263. 88 Bishir, J. and Namkoong, G . 1987. Unsound seeds in conifers: estimation of numbers of lethal alleles and of magnitudes of effects associated with the maternal parent. Silvae Genet. 36: 180-185. Bollen, W . B . , L u , K . C . , Trappa, S . M . , and Tarrant, R .F . 1969. Influence of Sitka alder on soil formation and microbiological succession on a land slide of alpine origin on Mount Rainier. U . S . Dep. A g r i c , For. Serv. Pac. N W For. Range Exp. Sta., Res. Note 103: 1-5. Boltz, B . A . , Bongarten, B . C . , and Teskey, R .O. 1986. Seasonal patterns of net photosynthesis of loblolly pine from diverse sources. Can. J. For. Res. 16: 1063-1068. Bongarten, B . C . , and Teskey, R .O. 1986. Water relations of loblolly pine seedlings from diverse geographic origins. Tree Physiol. 1: 265-276. Borchert, R. 1991. Growth periodicity and dormancy. In Physiology of Trees. Edited by Raghavendra, A . S . John Wiley & Sons, N e w York. 509 pp. Bormann, B .T . 1983. Ecological implications of phytochrome-mediated seed germination in red alder. For. Sci . 29: 734-738. Brissette, J .C. and Barnes, B . V . 1984. Comparisons of phenology and growth of Michigan and western North American sources of Populus tremuloides. Can. J. For. Res. 14: 789-93. Campbell, R . K . and Ritland, S . M . 1982. Regulation of seed germination timing by moist chilling in western hemlock. N e w Phytol. 92: 173-182. Campbell, R . K . , and Rediske, J .H. 1966. Genetic variability of photosynthesis efficiency and dry-matter accumulation in seedling Douglas-fir. Silvae Genet. 15: 65-72. Cannell, M . G . R . and Willet S.C. 1976. Shoot growth phenology, dry matter distribution and root: shoot ratios of provenances of Populus trichocarpa, Picea sitchensis and Pinus contorta growing in Scotland. Silvae Genet. 25: 49-59. Cannell, M . G . R . , Murray, M . B . and Sheppard, L . J . 1987. Frost hardiness of red alder (Alnus rubra) provenances in Britain. Forestry (Oxf) : 60: 57-67. Chaisurisri, K . , Edwards, D . G . W . and El-Kassaby, Y . A . 1992. Genetic control of seed size and germination in Sitka spruce. Silvae Genet. 41: 348-355. Charlesworth, D . 1989. Why do plants produce so many more ovules than seeds? Nature 338: 21-22. Ching, T . M . 1959. Activation of germination in Douglas-fir seed by hydrogen peroxide. Plant Physiol. 34: 557-563. 89 Chomba, B . M . , Guy, R . D . and Weger, H . G . 1993. Carbohydrate reserve accumulation and depletion in Engelmann spruce (Picea engelmannii Parry): effects of cold storage and pre-storage CO2 enrichment. Tree physiol. 13: 351-364. Cline, R . G . , and Campbell, G.S. 1976. Seasonal and diurnal water relations of selected forest species. Ecology 57: 367-373. Climatic atlas Canada. 1984. 1. Temperature and degree days. Canadian Government Publishing Centre, Ottawa. Condon, A . G . , Richards, R . A . , and Farquhar, G .D . 1987. Carbon isotope discrimination is positively correlated with grain yield and dry matter production in field-grown wheat. Crop Sci. 27: 996-1001. Czabator, F.J . 1962. Germination value: an index combining speed and completeness of pine seed germination. For. Sci. 8: 386-396. Dang, Q.L. , X i e , C . Y . , Y i n g , C , and Guy, R . D . 1994. Genetic variation of ecophysiological traits in red alder (Alnus rubra Bong.). Can. J. For. Res. 24: 2150-2156. Davidson, R . H . , Edwards, D . G . W . , Sziklai , O. and El-Kassaby, Y . A . 1996. Genetic variation in germination parameters among populations of Pacific silver Fir. Silvae Genet. 45: 165-171. Deans, J.D. and Harvey, F.J . 1996. Frost hardiness of 16 European provenances of sessile oak growing in Scotland. Forestry (Oxf) : 69: 5-11. Di l lon , W . R . and Goldstein, M . 1984. Multivariate Analysis: Methods and Applications. John Wiley & Sons, N e w York. 587 pp. Dunlap, J . M . , Braatne, J .H. , Hinckley, T . M . , and Stettler, R.F. 1993. Intraspecific variation in photosynthetic traits of Populus trichocarpa. Can. J. Bot. 71: 1304-1311. Dunlap, J . M . , Heilman, P.E. , and Stettler, R .F . 1994. Genetic variation and productivity of Populus trichocarpa and its hybrids. Can. J. For. 24: 1539-1549. Ehleringer, J.R., Ha l l , A . E . and Farquhar, G .D . 1993. Stable Isotopes and Plant Carbon-Water Relations. Academic Press, San Diego. 555 pp. Eiche, V . , and Andersson, E . 1974. Survival and growth in Scotch pine (Pinus sylvestris L.) . Provenance experiments in Northern Sweden. Theor. App l . Genet. 44: 49-57. Eiga, S., and Sakai, A . 1984. Altitudinal variation in freezing resistance of Saghalien fir (Abies sachalinensis). Can. J. Bot. 62: 156-160. 90 El-Kassaby, Y . A . , and Park, Y . S . 1993. Genetic variation and correlation in growth, biomass, and phenology of Douglas-fir diallel progeny at different spacings. Silvae Genet. 42: 289-297. El-Kassaby,Y.A. , Edwards, D . G . W . , and Taylor, D . W . 1992. Genetic control of germination parameters in Douglas-fir and its importance for domestication. Silvae Genet. 41: 48-54. Elliott, D . M . and Taylor, I.E.P. 1981. Germination of red alder (Alnus rubra) seed from several locations in its natural range. Can. J. For. Res. 11: 517-521. Erstad, J .L.F. 1994. Adaptation of woody plants to weather and climate in a maritime environment. I. hardiness. Nor. Landbrforsk. 8: 235-248. Falconer, D.S . 1989. Introduction to Quantitative Genetics. Longman Inc., New York. 438 pp. Farquhar, G .D . and Sharkey, T .D. 1982. Stomatal conductance and photosynthesis. Ann. Rev. Plant Physiol. 33: 317-45. Farrar, J .L. 1995. Trees in Canada. Fitzhenry and Whiteside Ltd., Markham. 302 pp. Fites, J .A. , and Teskey, R .O. 1988. CO2 and water vapour exchange of Pinus taeda in relation to stomatal behavior: test of an optimization hypothesis. Can. J. For. Res. 18: 150-157. Flint, H . L . 1972. Cold hardiness of twigs of Quercus rubra L . as a function of geographic origin. Ecology 53: 1163-1170. Fowells, H . A . 1965. Silvics of forest trees of the U S A . U .S . Dep. A g r i c , For. Serv., Agric. Handb. N o 271.762 pp. Fuchigami, L . H . , Weiser, C.J . , Kobayashi, K . , Timmis, R. and Gusta, L . V . 1982. A degree growth stage (°G) model and cold acclimation in temperate woody plants. In Plant cold hardiness and freezing stress. V o l . 2. Edited by L i , P . H . and Sakai, A . Academic Press, Orlando, Fla. 694 pp. Gallo, L . A . 1985. Genetic and environmental variation in aspen. Part 1. Germination and weight of seed. Silvae Genet. 34: 171-181. Giannini, R. and Bellari , C . 1995. Heritability estimate of seed germination parameters in Pinus leucodermis Antoine. Seed Sci. Technol. 23: 385-392. Gower, S.T., Reich, P .B . and Son, Y . 1993. Canopy dynamics and aboveground production of five tree species with different leaf longevities. Tree Physiol. 12: 327-345. Gross, K . and Hettesheimer, W . 1983. Comparative gas exchange measurements on fast and slow growing 18-year-old spruce of various geographical provenances. A l l g . Forst. Jagdztg. 154: 133-139. 91 Gutterman, Y . 1992. Maternal Effects on Seeds during Development. In S E E D S , The Ecology of Regeneration in Plant communities. Edited by Fenner, M . C A B international. 373 pp. Haeussler, S. and Tappeiner, J.C. 1993. Effect of the light environment on seed germination of red alder (Alnus rubra). Can. J. For. Res. 23: 1487-1491. Haeussler, S., Coates, D . and Mather, J. 1990. Autecology of Common Plants in British Columbia: a Literature Review. Forestry Canada, F R D A report 0835-0752. 272 pp. Harrington, C . A . , and Deal, R . L . 1982. Sitka alder, a candidate for mixed stands. Can. J. For. Res. 12:108-111. Heide, O . M . 1993. Daylength and thermal time responses of budburst during dormancy release in some northern deciduous trees. Physiol. Plant. 88: 531-540. Heide, O . M . , Junttila, O. and Samuelsen, R.T. 1976. Seed germination and bolting in red beet as affected by parent plant environment. Physiol. Plant. 36: 343-349. Holowachuk, D . L . 1993. Population differences in water-use efficiency for Pinus contorta Dougl. as indicated by stable carbon isotopic composition. M . S c . thesis, University of British Columbia. International Seed Testing Association: International rules for seed testing. 1985. Seed Sci. Technol. 13: 299-513. Johnsen, K . H . , Seiler, J.R. and Major, J.E. 1996. Growth, shoot phenology and physiology of diverse seed sources of black spruce: II. 23-year-old field trees. Tree Physiol. 16: 375-380. Joyce, D . G . 1987. Adaptive variation in cold hardiness of eastern larch, Larix laricina, in northern Ontario. Can. J. For. Res. 18: 85-89. Kenady, R . M . 1977. Regeneration of Red Alder. In Utilization and Management of Alder. Edited by Briggs, D . G . , De Be l l , D.S. and Atkinson,W.A. U .S . Dep. A g r i c , For. Serv. Pac. N W For. Range Exp. Sta., Portland, Oregon. Gen. Tech. Rep. PNW-70 . 379 pp. Keyes, M . R . , and Grier, 1981. Above- and below-ground net production in 40-year-old Douglas-fir stands on low and high productivity sites. Can. J. For. Res 11: 599-605. Kolb , P.F. and Robberecht, R. 1996. High temperature and drought stress effects on survival of Pinus ponderosa seedlings. Tree Physiol. 16: 665-672. Kozlowski , T .D. and Pallardy, S.G. 1997'a. Physiology of Woody Plants. Academic Press, San Diego. 411 pp. 92 Kozlowski , T.T. and Pallardy, S.G. 1997b. Growth Control in Woody Plants. Academic Press, San Diego. 641 pp. Krajina, V . J . , K l inka , K . , and Worrall, J. 1982. Distribution and ecological characteristics of trees and shrubs of British Columbia. University of British Columbia, Faculty of Forestry, Vancouver, B . C . 131 pp. Kramer, P.J. and Kozlowski , T .D. 1979. Physiology of Woody Plants. Academic Press, Orlando. 811 pp. Kriebel, H . B . and Wang, C . 1962. The interaction between provenance and degree of chilling in bud break of sugar maple. Silvae Genet. 11: 125-130. Kriebel, H . B . , Bagley, W.T. , Deneke, F . H . , Funisch, R . W . , Roth, P., Jokela, J.J., Merritt, C , Wright, J .W. and Will iams, R . D . 1976. Geographic variation in Quercus rubra in north central United States plantations. Silvae Genet. 25: 118-22. Kuzina, G . V . and Kalinina, G . A . 1993. Abscisic acid level, dormancy induction, and frost resistance of apricot plants in relation to photoperiodic conditions in autumn. Sov. Plant Physiol. 40: 345-352. Larsen, J .B. and Mekic , F. 1991. The geographic variation in European silver fir (Abies alba M i l l . ) . Silvae Genet. 40: 188-198. Ledig, F.T. and Kurbobo, D.R. 1983. Adaptation of sugar maple populations along altitudinal gradients: photosynthesis, respiration and specific leaf weight. A m . J. Bot. 70: 256-265. Leinonen, I . 1996. Dependence of dormancy release on temperature in different origins of Pinus sylvestris and Betulapendula seedlings. Scand. J. For. Res. 11: 122-128. Leverenz, J .W. 1981. Photosynthesis and transpiration in large forest-grown Douglas-fir: diurnal variation. Can. J. Bot. 59: 349-356. Logan, K . T . 1971. Monthly variations in photosynthetic rate of j ack pine provenances in relation to their height. Can. J. For. Res. 1: 256-261. Ludlow, M . M . and Jarvis, P .G . 1971. Photosynthesis in Sitka spruce (Picea sitchensis (Bong.) Carr.) I. General characteristics. J. App l . Ecol . 8: 925-953. Luoma, S. 1997. Geographical pattern in photosynthetic light response of Pinus sylvestris in Europe. Funct. Ecol . 11: 273-281. Lyons, C P . 1952. Trees, Shrubs and Flowers to K n o w in British Columbia. J . M . Dent and Sons Ltd. , Vancouver. 168 pp. 93 M a , Chang-Geng. 1989. Geographic variation Pinus armanii Franch. Silvae Genet. 38: 81-89. McVean , D . 1955. Ecology of Alnus glutinosa (L.) Gaertn. II. Seed distribution and germination. J. Ecol . 43: 61-71. Mebrahtu, T. and Hanover, J .W. 1991. Family variation in gas exchange, growth and leaf traits of black locust half-sib families. Tree Physiol. 8: 185-193. Mohn, C . A . , and Pauley, S. 1969. Early performance of cottonwood seed sources in Minnesota. Minn . For. Res. Serv. Notes 207: 1-4. Morgenstern, E . K . 1969. Genetic Variation in seedlings of Picea mariana (Mi l l . ) B S P . I. Correlation with ecological factors. II. Variation patterns. Silvae Genet. 18: 151-67. Morgenstern, E . K . 1996. Geographic Variation in Forest Trees: Genetic Basis and Application of Knowledge in Silviculture. U B C Press, Vancouver. 209 pp. Myking, T. and Heide, O . M . 1995. Dormancy release and chilling requirement of buds of latitudinal ecotypes of Betulapendula and B. pubescens. Tree Physiol. 15: 697-704. Nienstaedt, H . 1967. Chi l l ing requirements in seven Picea species. Silvae Genet. 16: 65-68. Nooden, L . D . , Blakley, K . A . and Grzybowski, J . M . 1985. Control of seed coat thickness and permeability in soybean; a possible adaptation to stress. Plant Physiol. 79:543-545. Orozco Segovia, A . , Sanchez Coronado, M . E . and Vazquez Yanes, C.1993. Effect of maternal light environment on seed germination in Piper auritum. Funct. Ecol . 7: 395-402. Owens, J .N. 1991. Flowering and Seed Set. In Physiology of Trees. Edited by Raghavendra, A . S . John Wiley & Sons, New York. 509 pp. Park, Y . S . and Fowler, D.P . 1984. Inbreeding in black spruce (Picea mariana (Mi l l . ) B.S.P.): self fertility, genetic load, and performance. Can. J. For. Res. 14: 17-21. Pulkinnen, P. 1993. Frost hardiness development and lignification of young Norway spruce seedlings of southern and northern Finnish origin. Silva Fenn. 27: 47-54. Radwan, M . A . and DeBel l , D.S. 1981. Germination of red alder seed. U . S . Dep. A g r i c , For. Serv. Pac. N W For. Range Exp. Sta., Portland, Oregon. Research Note N o . PNW-370. 4 pp. Rehfeldt, G . E . 1980. Cold acclimation in populations of Pinus contorta from the northern Rocky Mountains. Bot. Gaz. 141: 458-463. 94 Rehfeldt, G . E . 1987. Adaptive variation and seed transfer for ponderosa pine in central Idaho. U . S . Dep. A g r i c , For. Serv. Intermount. Res. Sta. Research Note N o . INT-373. 6 pp. Rehfeldt, G . E . 1988. Ecological genetics of Pinus contorta from the Rocky Mountains (USA) : a synthesis. Silvae Genet. 37: 131-135. Reich, P .B. , Koike , T., Gower, S.T. and Schoettle, A . W . 1995. Causes and consequences of variation in conifer leaf life span. In Ecophysiology o f Coniferous Forests. Edited by Smith, W . K . and Hinckley, T . M . Academic Press, San Diego. 338 pp. Reich, P .B. , Walters, M . B . and Ellsworth, D.S. 1992. Leaf life-span in relation to leaf, plant, and stand characteristics among diverse ecosystems. Eco l . Monogr. 62: 365-392. Ruel, J.J., Ayres, M . P . and Korpilahti, E . 1996. Variation in temperature responses among populations of Betulapapyrifera. Si lva Fenn. 30: 145-158. Sakai A . , and Weiser, C.J . 1973. Freezing resistance of trees in North America with reference to tree regions. Ecology 54: 118-126. Sakai, A . 1983. Comparative study on freezing resistance of conifers with special reference to cold adaptation and its evolutive aspects. Can. J. Bot. 61: 2323-2332. S A S . 1988. SAS/STAT® User's Guide, Release 6.03 Edition. Cary, N . C . : S A S Institute Inc. For analysis used S A S version 6.12. 1028 pp. Schalin, I. 1968. Germination Analysis of Gray Alder {Alnus incand) and Black Alder (Alnus glutinosd) seeds. In Biology of Alder. Edited by Trappe, J . M . , Franklin, J.F., Tarrant, R .F . and Hansen, G . M . U .S . Dep. A g r i c , For. Serv. Pac. N W For. Range Exp. Sta., Portland, Oregon. 292 pp. Schultz, R . C . and Gatherum, G . E . 1971. Photosynthesis and distribution of assimilate of Scotch pine seedlings in relation to soil moisture and provenance. Bot. Gaz. 132: 91-96. Seif, S. and Gruppe, W . 1985. Chi l l ing requirements o f sweet cherries (Prunus avium) and interspecific cherry hybrids (Prunus X spp.). Acta Hortic. 169: 289-294. Shulze, E . , and Ha l l , A . E . 1982. Stomatal responses, water loss and CO2 assimilation rates of plants in contrasting environments. In Physiological Plant Ecology: II. Encyclopedia of Plant Physiology, V o l . 12B, N e w Series. Springer-Verlag, Berlin. SigmaStat. 1994. User's Manual. Jandel Scientific Software, San Rafael, California. Si l im, S.N. and Lavender, D.P. 1994. Seasonal patterns and environmental regulation of frost hardiness in shoots of seedlings of Thuja plicata, Chamaecyparis nootkatensis, and Can. J. Bot. 72:309-316. 9 5 Smith, J . W . M . 1976. Nutritional effects on glass-house lettuce seed parent plants and their progeny. Hort. Res. 16: 45-52. Smithberg, M . H . , and Weiser C.J . 1968. Patterns of variation among climatic races of red osier dogwood. Ecology 49: 495-505. Sorensen, F . C . and Cress, D . W . 1994. Effects of sib mating on cone and seed traits in coastal Douglas-fir. Silvae Genet. 43: 338-345. Sorensen, F .C . and Ferrell, W . K . 1973. Photosynthesis and growth of Douglas-Fir seedlings when grown in different environments. Can. J. Bot. 51: 1689-1698. Stahl E . G . and Persson, B . 1992. Provenance variation in early growth and development of Picea mariana (Mi l l ) B.S.P. Stud. For. Suec. 187:1-16. Steiner, K . C . 1979. Patterns of variation in bud burst timing among populations in several pine species. Silvae Genet. 28: 185-94. Sun, Z.J . , Livingston, N . J . , Guy, R . D . and Ethier, G.J . 1996. Stable carbon isotopes as indicators of increased water use efficiency and productivity in white spruce {Picea glauca (Moench) Voss) seedlings. Plant Cel l Environ. 19: 887-894. Sundblad, L . G . , and Andersson, B . 1995. N o difference in frost hardiness between high and low altitude Pinus sylvestris (L.) offspring. Scand. J. For. Res. 10: 22-26. Sweet, G .B . , and Wareing, P.F. 1968. A comparison of the rates of growth and photosynthesis in first-year seedlings of four provenances of Pinus contorta. Ann . Bot. 32: 735-751. S Y S T A T . 1996. S Y S T A T 6.01. SPSS Inc. Tanaka, Y . , Brotherton, P.J., Dobkowski, A . and Cameron, P .C. 1991. Germination of stratified and non-stratified seeds of red alder at two germination temperatures. New For. 5: 67-75. Thomson, A . J . , and El-Kassaby, Y . A . 1993. Interpretation of seed-germination parameters. New For. 7: 123-132. Tognetti, R., Michelozzi , M . and Giovannelli, A . 1997. Geographical variation in water relations, hydraulic architecture and terpene composition of Aleppo pine seedlings from Italian provenances. Tree Physiol. 17: 241-250. Tremblay, M . F . and Lalonde, M . 1987. Effect of photoperiod and temperature on the development of frost hardiness in three Alnus species. Physiol. Plant. 70: 327-331. Viereck L . A , and Little, E . L . 1972. Alaska trees and shrubs. U . S . Dept. Agric . For. Serv., Handb .No410 . 9 6 von Caemmerer, S., and Farquhar, G .D . 1981. Some relationships between the biochemistry of photosynthesis and gas exchange of leaves. Planta 153: 376-387. Weber, J.C. and Sorensen, F .C . 1992. Geographic variation in speed of seed germination in central Oregon ponderosa pine (Pinusponderosa Dougl. ex Laws). U .S . Dep. A g r i c , For. Serv. Pac. N W For. Range Exp. Sta., Portland, Oregon. Res. Pap. P N W - R P 444. Weins, D . , Calvin, C . L . , Wilson, C . A . , Davern, C L , Frank, D . and Seavey, S.R. 1987. Reproductive success, spontaneous embryo abortion, and genetic load in flowering plants. Oecologia71: 501-509. Will iams, G.J . , Pellett, N . E . and Kle in , R . M . 1972. Phytochrome control of growth cessation and initiation of cold acclimation in selected woody plants. Plant Physiol. 50: 262-265. Will iams, M . W . 1984. The cold hardiness adaptive response of green ash to geoclimate gradients. Ph.D. Thesis, Perm. State Univ . In Sakai, A . , and Larcher, W . 1987. Frost Survival of Plants. Responses and Adaptation to Freezing Stress. Springer-Verlag, Berlin. 321 pp. Wright, J .W., Kung, F . H . , Read, R . A . , Lemien, W . A . , and Bright, J .N. 1971. Genetic variation in Rocky Mountain Douglas-fir. Silvae Genet. 20: 54-60. Wright, R . D . 1971. Local photosynthetic ecotypes in Pinus attenuata as related to altitude. Madrono 21: 254-257. Ying , C C 1991. Performance of lodgepole pine provenances at sites in southwestern British Columbia. Silvae Genet. 40: 215-23. Ying , C C and Bagley, W.T . 1976. Genetic variation of eastern cottonwood in an eastern Nebraska provenance study. Silvae Genet. 25: 67-73. Zasada, J . C , Norum, A . R . , V a n Veldhuizen, R . H . and Teutsch, C E . 1983. Artificial regeneration of trees and tall shrubs in experimentally burned upland black spruce/feather moss stands in Alaska (Picea mariana). Can. J. For. Res. 13: 903-913. Zelawski, W . and Goral, I. 1966. Seasonal changes in the photosynthetic rate o f Scotch pine seedlings grown from various provenances. Acta Soc. Bot. Pol . 35: 587-598. Zhang, J., Marshall, J .D. and Jaquish, B . C . 1993. Genetic differentiation in carbon isotope discrimination and gas exchange in Pseudotsuga menziesii. Oecologia 93: 80-87. A P P E N D I X 1 A N O V A table for germination parameters. P = population, E = error. variable source SS %SS df F P>F variance % variance EMS R'soT P 0.0298 95.36 26 64 0.0001 0.0003 94.03 CT2E +4CT2P Error 0.0014 4.64 81 0.0000 5.97 Total 0.0312 107 0.0003 G C T P 9.1123 96.77 26 93.28 0.0001 0.0867 95.85 o2E +4a 2 P Error 0.3043 3.23 81 0.0038 4.15 Total 9.4166 107 0.0904 G V T P 254.6278 97.36 26 114.81 0.0001 2.4270 96.60 a 2 E +4a2p Error 6.9093 2.64 81 0.0853 3.40 o-2E Total 261.5372 107 2.5123 G U P 21019.6166 83.10 26 15.32 0.0001 188.9181 78.16 a 2 E +4CT2P Error 4274.7362 16.90 81 52.7745 21.84 a 2 E Total 25294.3528 107 241.6926 G U T P 5.9661 82.09 26 14.28 0.0001 0.0533 76.86 a 2 E +4a 2 P Error 1.3012 17.91 81 0.0161 23.14 a 2 E Total 7.2673 107 0.0694 9 8 APPENDIX 2 Factor analysis for germination parameters Latent Roots (Eigenvalues) of 5 p r i n c i p a l components 1 2 3 4 5 Component loadings 3.499 1.349 0.119 0.033 0.000 1 2 PV 0.965 0.231 R'50T -0.694 • 0.675 GCT • 0.879 0.459 GVT 0.943 0.331 GUT 0.651 -0.721 Variance Explained by Components 1 2 3.499 1.349 Percent of T o t a l Variance Explained 1 2 69.974 26.980 Rotated Loading M a t r i x ( VARIMAX, Gamma = 1.0000) 1 2 PV 0.938 0.323 R'50T -0.224 -0.942 GCT 0.988 0.084 GVT 0.974 0.226 GUT 0.162 0.958 "Variance" Explained by Rotated Components 1 2 2.881 1.967 Percent of T o t a l Variance Explained 1 2 57.613 39.342 C o e f f i c i e n t s f o r Standardized Factor Scores 1 2 PV 0.324 0.004 R'50T ' 0.101 -0.529 GCT 0.394 -0.152 GVT 0.359 -0.063 GUT -0.130 0.551 A P P E N D I X 3 99 3 • l-H > c o Mi u ox d o jg *u fc o o 13 o '8 o u O to fa CM (U O rH A to XI o CD tO rH rH •H P-l o XI P. CM r- O CD to (0 O CO ro CD -H > c A ro rH rH tu CM P r~ LO CM CO rH o CM CD VQ CO o CO CO ON Q rH ro o LO rH ON r~ C T> (0 -H in L  P ON > 1—1 CM CM r- CM O CD o p •SJ1 to CO CN CD Oi o o o O C EH rH CO LO ro •H p -rH to CM CD CO Xi CD CO rH < 4-1 ro C rH CO r - CM O 4-> Q (0 o O o o (0 o CD CO CM rH O 0 o o o 1 1 1 C a P C P CO •H l-H -H CO tO u o o o o -H CD [IJ CM CO o rH to c e U o N Q rH r— CM c o M CJ CO ON o 0 fa o a CD C CD C M< CO •rH a CM ro 0 P 0) co LO P o •H to Q • ro E Tl Ti P o o e 3 P ro CO rH CM a ON CM rH r- to 2 CO LO -H -H O r- o ro <rj< m X P ON in r— ON rH o X CD H r- r- CO o 0 m u rH CO co •^ r CD rH o xi EH LO CM r~ o — P. Ti o CD rH rH -=r u rH u p < 4-> c u CD r- ON U 0 a CJ o o o o to rC u O rH rH rH 0 I P fa o CM co CM a CO Ti O 1 (0 4-1 w 0 Q CM rH < CO CD c H-l o CO O O o O P CO fa o P tO u rH to e fa II CO 3 to r C 2 -P CO > o P Z Ti rH O CD Q) •H a 3 rH rH c C H ro P XI CD rH o ON CO ^  0 rH 4-i tO CM LO CD rH Ti ro -H CO CO ro r- c ro to o CO •H O CO LO ro CP CD 0 -p rH ro o o to o II P 1—1 CO o CD c 4-1 -H m CO rH •a1 o u Ti -H CM rH tO EH r- o rH CO 0 CO a r—1 LO CO r- CO c fa CO CM LO -P O ro > eC rH 3 0 <D [-- ro CM o CD (0 CO O CO CO CD CO o O o o O o •m a u x; X o rH CO CO •H 3 LO -H c o I 1 4-1 T3 to u o o O l EH s O CO CO o 4-1 rH ON c 0 HH to < cj o o P ro to rH o •H rH o rH rH o o o II > rH c 4-1 o CO CO CM ro ro X P < • u o c c CD O o o CM CD -p rH -H c 0 p ro rH CM co rH r~ 4-1 rH o rH CTl CM CM u •H CD O ON r- CO ro ftJ to -H rH rH LO r- p 3 TS CM rH CM P Xi 1—1 CD o •5J1 CO 0 O 4-> ON O O CO CO 0 0 o rH ON CM ro P EH rH ON rH ro 0 -H to ro VO ro o CD 0 o ro CTl LO CM > < H c H rH CM LO ON EH r C -H LO rH CM •H c u o o o o 0 CD 00 LO ro O -H H-l ON ON rH r- P CD o 1 C c rH rH ro rH r- CO rH rH CD Pi CD (0 rH \0 o o O CD cC rH CM CD ON 3 CD o o .Y CM co CO ON S 4-1 o -H • CD 4-1 PI o o O O XI XI (0 CO T3 c EH a rH CM ro •=rf X2 O CO > CJ •H rH CM ro g •H < EH fa Z X! 0 ro P o < PI O to -H ro CJ HH H PI c P rH 0 CO - CD •H -H CO U 4-) P U rO CO rH o rH P -H CJ CD CO OS A P P E N D I X 3 100 to CD CD c O OO LO CO rH > 0 o 00 CX IX X! CO •H •H CM oo co CO rO CD 4-> 4-1 CM CM CM CM -H rH (0 H U X) rH o O O O o rO LO CD rO > CX rH rH t— r- r- oo oo CM 4-1 -H 6 o 2 rH 00 L O O <rj* rH •^p O LO •H M u rH Pi <x rH CO rH L O M rH O oo rH O CO (0 O ro [J rH o 00 rH o LH o o o o o O > o O < o a •H o O o o o CJ o o o o to OH rH c c CO O CD O ro o (0 CD rXl rH 0 c O OO IX oo c rH X5 CD -H o O oo CM o ro to XI rO X! C •H CM rH oo o 0 le fO -H -H U EH 0 C rt CM o o o u X! r4 rO rfl o O O o o •H to CO CM > O CO oo CO CM CD CO 00 -H > 00 r- LO CO o x: 2 o rH CO •^p r— H O CM LO CM CO oo C n +J OO CO CO IX rH fO c H O O r- CM LO CO O 0-. W CM LO 00 r- > o EH rH CM rH rH CM -H •H T3 CD . •H CO 4-1 C O o o o o 4-1 CJ O O o O O >i ro (0 l OT ca 1 al C •H rH Xi LO r~ XI 00 CO 4-> 0 c 6 rO CD CM CD CO CD fO rH < U 0 U co r- CM o rH C CD >i •H ro CO CM rH o XJ •H CD >i Oi XI c 3 ro e x: o o cr o o O o •H u 4-1 t— c <D xs c w U CM CO in o CD CM CO r- 00 (0 X! CD (0 i m 2 CTi <Xi LO rH 0> 4H O CM CD o CD r- X! 4-1 G CJ oi > Pi rH CX oo oo CO 0 M rH 00 LO r- C •H W CM o o 00 CM Q) EH rH O o rH rH L H ro c CD X! CO < . . . XI 0 rH 0 O o o o O +J CD CJ o O o o o CD a •H 1 rH O 1 oi CD X 4-1 C X! >x> O H CD c o CM CM to CD fO rH C > o CM CM CM C CD •H (0 rO CO -H -H ro CO 00 •H u 0 •H CD 4-1 4-1 00 00 IX XI g 4-1 fO •H M rH ro M M CD > c rO XI rH 0 o o o o CD XI CM o > rO 0 OH Ol rH r- CO •vT oo LO H rH 00 rH LO c C -H g o 2 r- 1X1 r> CM CO fO O rH rH o oo rt) Xi 2 U H 0) Oi LO CO rH LO OO c 0 M LO CM L O rH CJ CD O ro CJ OJ sz U LO r— LO LO 00 O •H EH CO 00 N > 4-1 CD •H a < o • •H M o o O O O 4-1 0 U o o o O Xi •H rH c 1 fO c O 1 U CD fO 0) rH CO hJ rfl x: o C o 00 OO o CD 0) U Xi EH -H 0 CM o rH CO M C c •H ro LO X> 00 4-1 M CD CO 0 4-1 00 o CM o 0) 0 X! 4-1 c rH • X! O 4-1 cn (0 o o o O o u a cn o C o o in M 0 > U > D > O > D •H CM Oi CD CD a, oi CD CD e> 4-1 (0 rH CD rH CM oo M H O o A P P E N D I X 4 101 co >1 rH ro C to c o •H to CO cu SH Cn 0) SH-OT co cu c •H T> M (0 x: 4-> co o rJ fa o o LO CM I E H ft Oi W pa > o fa Xi ro -H iH-rtJ > 4-> C CD Tl C CU D H CD Q M O fa A XI O u C M a o rH CM CD -X Ti * O Oi S Ti i—1 M r- CO CD tO M LO o LO M -H CD o LO to 4-i a) CM rH o rH 3 U O H LO o o cr to >, CM rH W Oj E H CO r-o o ro <M M rH o o 0 ro ON CJ CD r-ro r-rH o — rH CO LO EH O CM LO CO r- — O o rH rH O o o A • o o o X) O rH rH u CTl ro D H CO ro ro t— CD CTl LO P -r - rH CD o o co CD 4-1 CO g •H 4-1 CO w o DH M CD 4-1 M CD 0 6 MH (0 rH H f0 C M LO ro CM l l 4H rH CM CD 0 0 tO * LO CM U TS rH CM 0 0 ON CD -H He r- CD M, 0 CM r-U 4-> OH rH 4-) (0 SH CT o ro 3 rH • CD Ti M O o ON TS to o O g C CJ CD ro CD CD OH m tO rH rH 0 H 4-1 CTl rH rH 0 tO CO CD o o H C M P H rH C rH CM CO o o CD M ro a> XI CO g •H 3 s 2 rH CD o rH rH D H Ti CD H-l CO r- ON CD H-l tO ON o 4-1 CD CD g ro 0 0 ON CO > g •H LO ro ro O tO 4-> o o o m g M CO * 0 CD (0 fa o o OH CM r- 1 i >i 0) rH M rH •a (0 X5 CD (0 M E H -H CD CO 3 M H-l tH < fa rH rH rH CO to C <, o a > fa hi o CD fa rH fa D H rH CM X! CJ CD tO OH E H 4-1 •H fa CO CO M, E H E H ft tO 2 <; o > M hH CJ fa ON A CD X! O 0 O M o C M O in E H hH fa CM >H CM OH CD O ON CD '—1 X! 0 0 tO CO -H -— [--rH CJ rH tO > rH 4-1 c CD Ti rH CM 0 0 C CD -X rH CD Ti •X rH D H 0 CC ro CD S a o o m rH CM 0 0 tO •X rH CD - H -X rH H H-> D H ro 3 rH Ti • 10 o CD C M U 0 H C M u c rH CD M CD XI CO g - H 3 2 CD Tl 4-1 CD CO > 0 H H g o CD oi > 1 CD u rH TS (0 X! CD f0 U g •H CD 3 M H-l E H CO f0 C < > fa rH D H CD 4-1 CO co CD -P tO g •H 4-1 CO fa U CD 4-1 CD g tO u to C M Ti rH CD tO U -H tO H-l 3 U cr to o EC Ti U tO TJ C tO -P CO CO C M E H U U 0 u X! 0 U C M rH CD 4-1 U CD o g H H tO rH frl tO C M U CD CD H-l 4-1 (0 CD g g -H (0 4-> M to f0 W C M fa Q X! tO -H rH CO > o CM CM CO ON CD o O LO O •^P rH CM O 0 0 ON rH 0 0 r- co -=r ro ON r~ rH LO LO O CD O ON CO LO r-CD CM 0 0 r -CD rH ON O CM I CM fa C J Oi H E H 2 1 A P P E N D I X 5 3 P <3 <u ^ la -DH DH ,—, w > ^ H^ >- b o ^ IC c3 ^ ^ ^ CN «T> A, + ^ C Q O? CQ DH >-l I-H t H rand > > > VO m rt LO © f-•^ f ON VO oi OO °° (N OO CN in r— _ , CN VO in O O ON —< © © m o © © CN - <—J o d d O TJ- ON (N oo r-» o in r-~ in in — CN CN O CN ^  UN vq CN OO l> ON l-" vo © In CN oo in rt VO o vq in in o fa ¥ > ST cd > cn CN id-ci c i + + o fcl <L> > P ? S3 s CN ^ VO >> fc Is s^ / —^' S3 S p p DH rt, S3 8 > § . CN -O - > DH ¥ > o S tl tl A A S3 S ^ + DH PH, ¥ cd > . CN CN > m pa DH o c i + + cd > rt vo CN in m c3 > CN CN pa ir> * DH P3 DH ¥ > -d-o r-ON vo CN 00 °°. CN 00 O ci CN + + o fcl CL) ¥ > + > vo in in -•dr ST P3 DH cd cd > > cd > vo m s fci fcl A V-- cd £ > > + 8? DH fe. ^ I > > CN OO VO m VO •<d- ON ^ CN CN + + rt-^ pa cd > o fcl CO ¥ > ffl ^H cd in rt o r~ ON VO CN 0O °°. CN OO CI > i> CN LT) CN ©ON CN cd > oo r- vo _ ^ S S S o ON 2 "•£> m . m oo vo • ci CN VO CN CN rt ^ CM rt CN CN © in" vo 3 * JQ cn vo rt CN CN in in oo CN oo cn cn o CN f- r— CN CN T—H o in oo ON r~ o o o o ON CN o ON cn o o o oo VO 1—1 cn O o o o O <—i o cn vo p o o d d d d d d d d in m ON 00 o ON VO cn ON cn CN oo oo rt VO cn VO o o •3- oo cn m CN O CN d^- cn •si- in O r—< cn o CN rn CN CN oo cn CN VO cn o o •sr o ON in ON © vo i — i in cn CN ON © o o o o © © © © o CN XT -3- © o p o o © © © © © © © O © © cn m d d d d © d © © d © © d © d d © d © © © o © © o © © © © f- © o d d © d © r—1 rt cn rt rt rt cn rt rt rt © © cn © o © 00 VO © o in © © © ON o © © m © © © o o © CN © © o CN © © © © © © © d © © © © d © © © © - V I CN TJ-cn oo ON • cn in CN P ON CN P cn cn TJ-cn cn oo i — i oo CN in ON cn f-; © ^r)-rn oi rH S in VO CN r - H . in" ^ «» m - H cn CN in m cn 00 cn in VO r~ CN m in cn in oo rt cn in vo CN in m cn ' 1 m 00 o in vo CN VO CN VO ON ON rt in ON VO VO © VO © m 00 00 t-y—t ON cn © oo ON vq cn •<t 00 vq ON vq m CN •d-cn ON © oo VO ON CN p © cn © cn CN in rt CN cn rt in 00° CN vo ci VO od d d vo VO d ON ^ r" Oi 00 VO 00 cn •dr cn 00 ON VO in ON cn cn r- CN cn cn CN f~- ON cn 00 ON f- m m CN in CN o oo 00 in CN CN © •n ON oo CN rt in VO in cn r-- cn © VO ON in in © rt ON VO ON 00 in © r~ CN ON vo 00 CN CN vq © ci vo rt o © o CN VO 00 ^r r-" ON CN ON © o © © © p © © d in CN •* CN ^ H CN d © © © © d © m in cn rt ON ON 00 CN CN vo rt ON in CN cn cn CN ON VO o rt rt m CN VO CN i—i VO VO f- o CN oo m cn in ON © CN VO ON © 00 ON VO cn ^ H VO VO m VO r- CN © CN VO 00 cn ON p~ ON vq cn © Tt" ON OO © cn © ON ON t-;00 00 © VO r^" cn CN © p~ oo CN rt cn" © CN cn m CN PQ DH PQ DH p cd fcl o pa DH CQ DH 1 +-* D Hl_( pa „ * P3 DH p cd tl O pa DH Pa DH P CS S 2 cd o u DH PQ DH i . w 2 ~£ A P P E N D I X 5 5 2 fc fc > C3 > m m * i. OH PH O fc ON ^ CN > co -i, + m S fc-(3 ca > > r- ON o fc tu o ca fc 5" S3 m £ fc = - > PH -St w CO s _; > in pa o fc > + CN OH > ca ON >• in ON CO * ca > •a-NO r— in sq oo •a-' d (N <N VN (N m t-» o Os ro in ^ S <N S 00 ^ ro ™ t-: ON ~ 1 ro m CN in oo o ON oo ON © o o o o o o o — i OO CN ON NO NO j^.' O —; Os ON -a" ^  CN -ST ro ro m ro CN ro -a;; © CN —I CN ro so CN ro CN in oo r- ON •* oo o in SO t~; in r~ SO OS -a-' ro 00 ON ro ro CN SO SO Os Os © O ON 00 CN in o ro ro O ro I-H CN CN CN ro ffl S S o ' S # c 5 fe PH pa cu 3 3 5 2 P 03 CQ _ PH « PH S 2 o fc (L) - 13 gP fc g fc| s ^ £ :> OS ON Kj» +^  m fc G, fc c3 cd > > > SO Os oo ro so ^ H O O in °} °^  ro m —< O o o O O P P P © o o o o so H^ SO CN in o o o in m m ON in 00 o o o ro in r^ SO ON ro o o o CN CN Os r^ ON ro o in © © o OS ON CN o O -a- o o o SO so d O H O *—1 CN o p p '—1 <—i d d d d d d d d d d so _ H 00 ro CN o o ro SO m >n o o so SO T-H o o •a- SO d in d d d d d >n r-oo in ON CN so (N so CN O in ON ON so r-CN in in ro CN ON CN CN CN -a-in OS © 00 ro sq SO p CN in ro C-; CN o r -ro so ro r~ in 00 m in (N ro in ON d CN (N ro CN CN ON ro od CN ro in CN CN CN d >n CN m oo ro in CN CO ro oo oo H^ r - r - ro © m r—i SO Os in oo CN oo ON r- H^ *o t m ON f~ oo in © SO ^ H oo m CN ON r~ © © •t SO CN ON r-; SO in in © »—i •a- CN r~ ON CN © so Os OO ON -n -H SO •a-' © PH pa PH S 2 H i Q PH > A P P E N D I X 6 104 CO -H CO >, rH (D C rO C O •H CO co CD U CP CD U CD CP C CO £1 O X CD co ro o X) (0 •H rH CO > 4-1 C CD X) a CD a CD a rH o 4-1 CD U P X i CD 0 O U C M CD CO •H a CD H-> CO <w o >! M CO 3 CO CO •H +J H (0 CM X ) rH H r ~ o X I rH M CM CD CO M CM CO CD (0 M CM fa m o rH •H rH •a1 fc. OO u -H CO fa ro o A CO ro CO 4-1 CD CD -=3< A CM cd 4-1 CD ^3" A o CM X! rH ro M a LO CO X! o p SH a LO X! o CO O o rH cr CO rH ro 0 o cr CO LO o o o S-l CO C M E H CT. LO M CO CM E H O U CM o o orr O o CM o C M o orr ro o C M o o fa LO o o fa LO o fa o a, o rH CM CD * X) * O Oi s u c CD M P 2 Oi CTl Ol rH CM ^ CD CM CTl CO ro r— rH r-LO rH CM CM ro rH o r-CM CM o o CO rH rH ro o r-CM o o o TS CD > O g CD cC CD rH X i Xt CD (0 U E H •H CD CO > rH 4-1 < Cd CO C O > Cd o Cd a, rH CM CD 4-1 CO o M C M X3 U M 0 CO rH X i rH C Cd CO 4-1 CO rH CD a) H-i 4-1 CD g ro u to ro u C M fa a xi ro •H rH CO > o r-ro co co o rH o oo LO r--LO LO LO rH LO CO ro co oo cn LO CM r- T co CM 0 0 rH CM o o o CM CM Ol o o o o o o cn -vr co •^J1 Oi CM CDrH O CM O CM O CM O I C M fa o CC Cd > EH CO Cd < Oi Ol cn Cd CD rH XJ ,—, r -ro •H — rH u o CM ro > rH 4-1 a CD X i rH CM LO c CD -K LO CD X i O a 0 CC ro CD 2 a O M O HH rH CM LO ro LO CD •H o H H-l ro P VH X i ro o CD CM 0 0 SH CM u c rH CD H CD X) CO •H p CD TJ 4-> CD CO > O HH g 0 CD CC > 1 CD u rH X i ro X! CD CO U g -H CD > P rH 4-1 fa CO ro C h i > Cd Cd a, rH CD 4-1 CO X! 0 S-l C M o II u CD 4-1 CD g MH ro SH EH ro C M x i SH u ro X3 C Cd ro 4-1 CO S-H CD CD 4-> 4-1 CO CD g g -H CO 4-1 SH co ro cd C M h-l o Cd CJ fa a X ) CO •H rH ro > O CM O O o o LO r-CO rH CTl 0 0 0 0 CM CTl CO CM r - r -O rH oo o r- o rH o o o r- •q* rH r-CM LO ro CD C M Cd o cC Cd EH 2 < CD rH X! CO •H U CO > 4-> c CD X i a CD 01 CD Q u o MH CD SH P X i CD 0 O SH C M CD CO -H CD 4-1 (XI m o SH CO CD X i o s X! g P 2 CD 4-1 CO CM LO <3< r- CTl CTl rH CM rH CM r~ rH cn CM •— rH CO CO CJ LO rH o ro ro CM CM CO CM r— CO o r~ CM LO o O o rH CM CM CD ro r- rH LO •H o r-4-1 CC o o M CO o o o CM M c rH CM ro CD hH X) CD > 0 g CD CC .CD rH X i X! CD ro U EH •H CD CO > S-l 4-1 < Cd E H ro C O hi < > fa u fa hi a rH CM 0 0 A P P E N D I X 6 105 TJ rH CD rd H -H rrj 4-> 3 M tr ro >i OT fa LH CD O J M M O CJ X ) rd •H H rd > LO co o co r- CM LO OO oo o LO LO rH oo ro r- oo LO CO CD CD O i—I CM O O O LO i—IT—I CO rt CM LO LO L O L O S T O fa A X) O u fa fa o rH O o CO A i * Cr o o O o X) CD O rH M XI rt LO O-i rO Oj rt LO •H —- OO r-u O LO r-ro • > CM rH o LO rH ON oo I CM CO rH 0 0 4-1 o U O rt rH o c fa CD CD 4-> CM rH CM 0 0 X! rH CM CO LO CO u CD c CD ro CO CD o g CD Xi + LO ro 4J 4H (0 O OJ CM ro CO U CD 2 g EH rO a o o •H O-i 4-1 u CO 0 fa X) LH rH CM CO CO U CO LO o rd ro rt U H 0 oo 0 0 rt LO CD -H * LO CO CD rd H CO LO r- LO H 4-1 0 J CM o 4-1 Xi rH oo o rH LO 3 H CD C fa rH OO LO rH •a rd o o g rd r> o O rH CD CL, rd 4-1 CO  o o o U OT CO o o o O rd » u fa CM o o o Oi u C rH CM CD M CD CO •H U CD CM CM CO rH CD 4J LO rt o rH a 4-> rO 0 0 0 0 0 0 LO CD •a CD e CO rH rH 0 0 4-1 CD g •H rt O o o OT > cd 4-1 0 0 O o o 0 H CO • . 4H g rd fa LO o o o 0 CD O-i OJ >1 CD u rH Xi rd XI CD g to U EH fa g -H CD OT > rH rH rH rH 3 (-1 4-1 < fa a OT ro C O trt > fa u fa 0 J fa CJ fa fa EH Z M EH > OT fa < fa o fa cj O J CD 4-» OT rt CM r- CM O OO O O O O CO O0 rt LO o r-LO o co oo Xi rH CD ro M -H CO 4-1 CD =1 u cr to OT 0 J O J > I EH U H O o XI o u 0 J M CD 4 J M CD O g m ro M EH rd 0 J o T3 r4 tO C fa rd 4-i OT r4 0) CD 4-1 4-) tO CD g e - H 10 4-1 M CO rO fa fa fa a XI rd •H H rd > co r- V rH M r- o CD ro H rH LO fa LO u •H O r- A o rd 4-1 CD LO ON XI o 3 H CH O0 LO o o cr ro >1 rH rH u OT OJ EH rH CM fa o SH O o M O fa rt CJ r-rt CM LO ON — o o  L O ON rH O LO O LO o O o g rH O • tr A o O o O CD XI rH O XI . rt U rO O J r- fa •H ' CM U CJ LO rd > oo CM rt CTi O CO ON LO r- 4J I rH oo [-- LO C o U r-' CD <D rt rH CM X) rH CM LO 4-> CO c CD * ON rH CD ro CD X3 0 0 O g 0 J 0 fa 0 0 LH rd CD S Q O &H rd fa M O L4H r—1 CM LO rt OO I— tO ON Xi U o r- LO O CD •H 0 0 SH 0 ro 0 0 r- CO H 4J fa 0 0 rd M o CO LO LO 3 M X) M CM LO ON r- X3 rd o C fa rH CM 0 0 r~- CD O J rO rH ON CO rt O 4-) o rt O rH 0 OT o • • U CM o o 0 J U c rH o 0 0 CD H CD X! CO g •H 3 0 0 rt o s Z M Q) LO rt LO r> a CD 4-1 r-0 0 CM o CD T3 4-1 tO 0 0 0 0 r- rH 4-1 CD CD g ro 0 0 rt 0 0 OT > g •H rt CM rH 0 0 0 rd 4-1 o 4H g M to CM O o O (1) rd fa o LO fa fa 0 0 > CD u rH TS ro XI CD g rd M EH •H CD OT rH rH rH 3 M +J < fa rH OT rO C O a > fa CJ 0 J CD fa fa rH fa CJ O J rH X! CJ OJ EH CD rd fa fa > OT 4J -H fa EH fa <, OT H EH Z o CO Z M fa o > H LO o o o CM r-rt o o o o o A P P E N D I X 6 106 cd •H CJ cd -P a CD TJ G CD a CD Q SH O 4H CD M D TJ CD U o CM CD CO -H g a CD 4-1 Kl 4H O > ra 3 co Tl rH M 00 fa CD rd H r~ o S4 •H rH A r- ra 4-1 CD O X! rH 3 H a CM 0 o cr ra >1 r-SH CO CM EH o CM o rr • CNJ o fa 0 CTl ro ro u LO EH rH O o r-CD — O rH a, u rH CNJ <D * Ti -x o cc S rd •* •H -x 4-1 CC u ra CM SH C CD H 3 z Tl CD > 0 g CD CC CD rH TJ X ! CD cd SH •H CD S-l 4-1 ra > CM CD 4-> CO CM LO CO ro CM l> O CM CM r-o CNJ > fa hi fa X! 0 U CM 4H cd U EH rd CM Ti U ra Ti C fa ra 4-1 CO SH CD CD 4-1 ra g H 4- 1 CD g ra 5- l CO ra fa CM fa a X! rd •H S-l ra > rH CD C O LO CD LO CNJ CM ON O LO LO C O LO CO LO CO C O o o o ro oo ON rH CNJ CD O CNJ CN] CM CO o • o o o CM fa o cC fa EH z 4-1 c CD Tl c OH 1 a? SH o 4H CD SH 3 Ti CD 0 O SH fa CD CO -H 4-1 CO 4H 0 ra Ti rH M LO CM fa CD ra H rH ro CM LO SH •H ro CM A CO ra 4-1 CD rH CO X! rH CM =1 SH CM CM 0 O rH cr ra > I LO LO M CO CM EH CO ON CM o O rr rH o o o fa 0 ON CM CM LO ro ro CJ rH LO EH rH o CM CO LO [-- CM — CM CM a, C J rH CM CD * Tl -X O CC s ra * cc SH C ai i—i X3 g 3 Z Ti CD > 0 g CD CC CD H TJ X! CD cd SH -H CD M 4-I ra c > fa DH CD 4-i CO ro r-r- CM ON CD •=p o CM CM r- o rH CD CM C3N CM CM O O r- ro rH CM r-CM O O O EH > CO fa <C hi O fa U X! 0 SH CM o SH DH CD 4-1 SH CD o g 4H id SH EH Cd Ti SH ra TJ a w ra 4-1 CO ra fa CM fa a XI ra •H SH ra > rH CD CM i ro ON ro ro LO CM rH rH I I CO LO LO CO CO CM CD o o o o o CO o o ro o o o o o o o o o o o rH CM 00 rH CO rH CD CM O O O • O O O O LO O o o o o o o I I CM fa u CC EH fa > CO EH W ft, Z J o H fa CJ A P P E N D I X 7 107 cn CO .2 > CD Oi cd .-=! o X CD CO cd at d o • i-H o o I—I o '3 o fa rH O A o to o CD U rH fa o Xi fa rH rH rt L£> rd o rH 0O rt •H a A O rH rH rt fa ro U rt LO oo CM CO o oo OO CO r— O o CN rt Q o rO O r~ ON OO LO Tl (0 •H ro CN O o M • • rH > H LO LO CM oo CD 0 4-1 CN 00 O CM CD fa o o O o C rH EH rH IO rH LO JH -H rd CD rH co o XI CD rH rH < m c rH oo CO r- r- 4-1 a rO u o O O o 3 o CD r- CO 00 rH O ON U o D1 c M C u LO -H fa OT rd H o O o O •H CD fa ro rt CO OO co c CJ O N a o LO oo rH c o CJ co rH CO o fa CO c C CD c rH CX> •H Q CM rO 0 U CD rH LO 4-1 U •H CO a • • (0 e 4-1 ON CO e 3 U ro rt rt LO r~ TJ oo rH LO CO rO LO rt •H z •H O LO LO LO oo X M LO CM r- oo rH 0 X CD M O r— rH LO 0 rO V4 LT) ro oo r— CD rH o x: LH o LO r- CM H TJ o OO LO CM CM M rH M 4-1 < • • a C u 00 CO CM LO M o TJ CJ o o o o a rd u o o rH rH 0 LH fa 00 CO o rt 9- LO O 1 4-> w . 0 a CM rH rH < fa rt CD C fa CO o o O O 4-1 I'- CM U rfl rH rO fa ll CM 3 ro si 3 z +J to O 4J Z TJ 00 0 CD -H c 3 M a c rH rd U XI rH o CO CO ON r- o rH 4-1 rfl CN CM rt oo LO Tl rd •H oo CN rt rH c rd CO rH OT -H O r- LO ON CM CD U 4-1 CM LO ON 00 rO 0 II H H CO LO CO ON +J •H rO r- ON CO CO O TJ •H £ CO H rd EH oo CM rH LO CO C rH VO CM LO C fa LO r~ O ro 4-1 LO rd > sC 3 O CD CO r- LO 00 CD rO rt r- 00 rt to CD CO 0 CJ o O o o -ro a U si X CN ON o r- •H 3 CM •H C o 1 TJ rd U o o O o EH O CM CN rt ON 4-> rH 00 a 0 fa < o O 0 r4 rfl rt rd LO O -H CJ M 00 CM rH o 4-1 II > OO C 4-1 o DH OT OT o rd ro fa 4-1 < CJ 0 C CD o O 4H CD 4-1 rH O SH rfl rH CO LO 00 CO rH o CN ON 00 rH r4 -H <u O CM o oo CM rd •H rt CM ON co 4-1 3 TJ CD r- co r- U X! M CM ON LO rH u 4-1 rt LO rH LO CO 0 o LO LO LO rt ro 4J EH O 00 LO CM •H rd 00 00 00 CM CD O 0 CD o CO 00 > < c rH oo rH rH rH EH X! -H CN o 00 ON •H C u O O o O o CD 00 CO CD rt •H 4-1 00 rH r-~ r- 4J CD o c M rH CO LO LO r- ON rH CD fa rd M o o O o CD fa 00 r> rH CM 3 "5 CJ 0 o rH LO 00 s 4-1 U •H • CD fa o o o o X5 ro to TS C EH rH CM 00 rt X) O OT > CJ rH CM 00 rt 6 -H ed EH fa z O rd 4-1 O fa o •H fa rfl CJ fa fa fa 44 rH CO - CD •r-l to U 4-1 M U rd rH 0 4-1 •H CJ OT s A P P E N D I X 7 108 co CD rH X) ro -H H ro > ro o - H c o c ro o SH • H <D x: 4-1 T S c ro to Q) rH XI ro - H SH ro > CD Cn C ro x; o x a) c CD <D s +J 0) X! OT C O •rH -P ro H (1) P P O T P H rf in H (31 00 T P rH T P L O Cs] T P CO fiC «) CTl O ( N I D H CD N u i to i n u i >j o o o o o o o T S C 10 ro oi rH to Xt o> ro rH -H X! P ro ro T P o H E H < u o PI O O T Ol CO N CM CTl T P Ol tO r - rH T P o T M H C N h co rH CM CM CSI CM rH rH o o o o o O O I ro C O rH T P cn C M rH T P •H > ro L O T P o rH so o T P C O r - O N ro C O ro ro C O P O T P r - O N T P C M L O rH < ro O N r - C O L O O N ro co C P I C O co C M L O ro C M O N C J rH L O cn o o co C O > OT AT r\ JL o ro C M o o T P C O o o o o o o o CD P o o o O o o o o Cn ro CM T P o CM ON T P as T P CO L O CNJ L O CM ro CO rH < T P T P r - T P CM Ol CM CD L O ro L O r - r - CM T P o o o o O O O ro rH CO CD T P CO Cs] CO cn CO LO CO LO CM ro o Cs] < T P •=3" CO r - T P T P r— tl) O CO ro rH ro ro rH rH XI H-l o o o o o o o OT Cn < ro u g -H Cn U C ro X! rH o X CD CD x; p CO ro 4H Cn 0 CD CO X i CD P rH X i c ro CD -P CD P s ro p > CD X! rH ro to o c - P o c •H O P c ro ro rH o CD P CD P X! 0 P u PI CM O M E H < u o PI r - CM O T P LO CM CO CO 00 CO T P CO ro CM T P r - CD o CO T P T P T P CM T P CD LO CM ro o O o O o O o E H < O o PI 00 r - CO r - CM CM T P CO rH L O CM r- L O T P o rH co ro o CM L O ro ro rH ro ro rH rH o o o o o o O w D 3 to CC Cd ON <c ro u 6 -H Cr> O Cd D S co >i rH ro c < o c ro T3 c 3 •a CD CC ro o •H c o c ro o CD c T P r - 00 o > O CD ro ON CM co -H -H CO CD CO CO CD p P o Cs] ro ro rH (0 P XI rH 0 o o o o CD to 3 a P -H g 0 •H p 3 p to to CJ a, O > a CO a rH CD o ro rH o c T P T P rH Cs] X! CD -rH 0 CO r— CD Cs] to x: c -H CD ON r- T P -H E H O P O rH o o P c P ro ro 0 O o o o > CJ do CD p Cn C M C ro X! O X rH T i CN 00 o CM CD tO CD 00 rH CD O 0 P ON SD r- r -co >i -p tO r - CD ro rH ro Xt c 3 Cn TS no Sq o O o o CD CD ro 1 X! C u CC tp o CD CD CD c Cs] T P CD ON O > o ro rH ro rH • C CO -H •p CO CO CO CO tO a> -P p o ro LO 00 -H rH ro p P XI rH o o o o o CO ro 3 a > wn -H p urn ro "0 O ro o C M CD N -H - P rH T l CD ro P X o c CM CM ro CM ro H •p 0 ro 00 CM CO V c -p 00 ON O T P c 0 p o Cs] CM CM ro c p p (0 o o o O O CO CJ a 0 p C M rH CNJ ro T P A P P E N D I X 8 109 HH N ? o S3 CJ cd > i-, ,—s cd i_ > s c oo cj co ^ t— oo co CN CO SO VO VO Ol rt CO vo CN CO vo VO CO ON —i CN CO rt ON O O 00 ON rt oo in ON -d-CN vo ON CN -^ r vo O cd ti o OH W H O 13 cj > + cd in > 2 ON i3 rt CJ ^ ¥ CO ^  CN > o !3 cj ¥ > cd !_* > 2 13 oo cj co « CN > o 13 cj t-< cd > a? > 2 CN S3 ON D °<= ^ CO ™ CN > o 13 CJ kH cd > ST ¥ > o S3 CJ ¥ > + cd > in CO * CN > O . S3 co <J ^ ¥ CO ?^ CN •> O S3 cj ¥ > * - , V cd > 2 >n C CO <U ¥ CO CN K > r - r - rt CN co vo vo vo CO f-co vq VO CO rt in CN vo i n rt O CO CO O 00 r - rt vo co co _ ; r— oo m CN co vo VO VO rt >0 CO VO C - rt 00 CN vo vo NO ON rt | -~ CN o ON in CN m C C N O N C N C O in VO VO 00 CN " O rt d ON rt <<0 ON o VO O rt rt ON in in 8 s DH W H in r -rt rt rt "1 ^ d rt 00 ON •>cr <d- m VO ON rt t - oo in ON rt rt C N VO ON oo r -oo oo in r -o S3 W CN rt rt ON VO V I CN CN m O cd fa o CO 00 ><o oo VO O CN -d" VO o S3 W t -vo o r - - H oo CN co m VO VO o o CN OO rt 00 in rt m rt m m vo ON UN M VO p CN CN 00 co ON d CN _1 < Q oo O cd S3 O DH W H H Q H r - rt oo CN co in vo vo VO rt co vq r-' CN CN r -ON rt rt CO ON CO ON in in rt o in vo CN oo o r - r-~ p vq vq rt CN co O cd S3 o W H o S3 CJ LH cd > + > 2 ON S3 rt CJ t- "cr CO * CN -> oo CN CN OO ON ON •d- VO CN 00 oo CN oo CN <—1 oo CO VO t-; CN O ON ©• ON CN r - © ON CN d ON in OO rt CN ON © d ON in CN t~ CO VO -d- in CO VO in CO VO m CN CO VO r- in CN t- in •n oo in CO ON CO CN in CN CO in ON CO VO in i ! in i—i r- CN in CN CO in o m 00 CN ON © © r - -d- CN ON ^ H m CO 00 rt o CO 1—1 oo ON in CN •n VO CN r— I- H ON co in C  • CN CO m ON CO 00 C  r—< •n r - vo "d- CN ON CN VO VO CO 00 vo VO 00 r - CN VO ON •d: oo CN CO © r—1 in VO m VO CO in CN ON 00 C  o •n oo CN ••d- CO o O © CN VO ON CO r~ p d ON vq ON •<t •d- 'd- oo CN CO vq vq vq CN o © © CN VO d d d d CN d d d d © © d © d d © © © © rt rt i — i rt rt rt o o P o © © © © o o o o © © © © o p o p © © o p d d d d © © © © CN VO 00 in ON CN oo ON CO CO ON in ON •n od in rt d m vd in CN CN CN CN r -C N CO vo VO VO © © OO CN rt m CO vo CN 00 rt "d" ON CN ON CN rt 00 CN © ~ © ON m ON rt CN -d-vo in d^-O cd S3 O DH W H < DH A P P E N D I X 8 o fc D In ca > -a-o o os so ro ro os r-so so so Os ro C N ro so C N ro C N ( N oo in ro OS CN CN os CN C OS SO SO r- ro so ro >-< oo so ro o ca fc z o fc ¥ > ST > o fc oo <D ro r— oo —< TT I/O ro so ro ro O OS oo oo oo O O O -a- ro ro t— SO OS © p d d o o o >/-> r--l> so ro CN OS CN CN 0O OS O so ro ro so CN ro >n OS so *—f ro r^  <n r- TT oo OS o Os 00 27.1 d t-: Os o o 1.44 OS in oo Os so *o 00 ro OS OS 1—1 1—1 a^-o <^  fc o a A P P E N D I X 9 i l l CO •H CO >i rH ra c rO C O • H CO CO CU p Cn CU P CO CP rH Xl ( 0 •H M ( 0 > tO O -H cn o rH o x: a P O £ rH CD T S rH O ac g d) CD rH rC 4-) to •H o P 4-1 ro C > -H 4-1 >i c SH 0 ) 4-1 TS c C CD 0 ) a M CD O a tp SH rH O CD m > CD CD rH SH 3 CD T S 0 CD C 0 ( 0 0 CJ SH •H DH 4 H •H CD C CO Cn •H •H CO a CD o H-l o CO LO rH o CD X! H-l 4-1 CD g CD rH X! tO •H rH rO > 0 2 TJ '—I M CD ( 0 H H P -H tO 4-> CD 3 SH cr to C O OH E H SH SH O c_> XI O SH D H O P X CD •p P CD O g m to p E H rO DH TS P tO TS C tO p C O P CD CD P P ( 0 CD g g -H tO P p CO tO W DH UH Q X! ( 0 • H P ro > o o o CTi CD P 0 r-r-rH C N r-CM r-o r-ro o r-D H w o OH U E H 2 M TS rH M r— TS rH HH rH CD ro Pt rH CD ro M ON CN P •H CO tu LO CM P - P rH A CM tO p CD CD A O CD ( 0 p CD 0 0 XI CM 3 p a rH Xt O rH 3 P a ro O O cr ro CM O O rH cr ro >1 CTl P CO O-i E H CTl P CO DH E H CM DH O orr rH O D J o o orr TT o CXJ CO o tu CO CTl CJ CTl O o a o rH CM 0 ) + T ) + O OH s CD CD CD CD O I CM CM ON rH CM CM ( 0 * CM • P -X CTl P C H rH P CO O DH P CH rH CD 1—I 3 2 Ti CD > 0 g CD OH CD rH Ti XI CD ro P E H -p 0 ) CO p H-l < ro C O > w CJ a rH CD P CO C M C M C M Tl CD CD LO r~ CD o O CM rH A . m o O X! CD 0 rH P X! . . ro TP CM ro a rH r-•H — CM ro P o CTl TP ro > rH rH O rH ON , I  r- ro 4-> o P CO TP c X Q) CD P rH CM Ti rH CM CO CTl p CD rH 1 C CD -X CD r~ o g CD Ti -X CO TP m CO a 0 OH ro TP P CD • 2 E H ( 0 Q o o CM P o tp rH CM CO rH Ti P CO o ro -X CD rH p 0 CO ro CD •H -X CO CD ro P CM TP P p OH CO O TS P TT CO 3 p C W TP CD TS ro o O tO CM o CD CM p CD o 0 CO rH o 0 • • P O o CM p C rH CM <D Pt CD o CO •H 3 p CD o CTl 5 2 CD 4-1 CO CD a P ro TP CD CD Ti CD g CO rH p CD g • P CM O CO > ( 0 4-1 ON O O P CO • tp g ( 0 u rH O 0 CD CM 1 OH > CD p rP Ti ro X! CD P ro P • P CD > tu rH rH 3 p 4-1 E H w Q CO ro C < PI > hi P I w CD CM rH w X! CJ a CM ro OH E H CD •p Cd CO P p E H < CO ro 2 o > Pt CJ XI o p C M O P X CD P P CD 0 g M H to P E H tO C M TJ P ro T> c w ro p CD P CD g ro p ro w C M UH Q XI CO • H P ro > ro C M O CD O rH O r H TP O O TP LO ro TP CM CD CD CM i—I CTl CD TP r H r- 0 0 TP r TP C M CO TP CO o o O CO CO ro TP CD i—I CM r- i— LO TP [ ro o ro CM LO rH o o o C M W u OH E H 2 A P P E N D I X 9 112 D H H CD ro M fa H i n u - H A LO CO (0 4-J CD X) o to s u a o o o tr ro >i SH • • C O fa E H OH O O SH SH O fa CO O CJ LD rt O ON r- — E H fa cn rt — CJ A CD XI rH O X! — ro rt SH CO r- CM fa -H — LO O SH CJ r~ ro CO > ro rH O 4-1 II a O SH CD fa CD T S rH CM rt rt 4-1 c CD -X CO rt SH CD CD TS -X r- CTl o g 0 fa CM ro LH rO CD s H a O o E H rO fa SH o fa rH CM rt o rfl -X ro rH T S SH CD •H -X r- CM u 0 H 4-1 fa CM rH rO U p H TS M TS CO o O C fa CD fa tO 0 4J 0 CO H fa M C rH CM CD M CD Q w -H 3: s-i CD a CD 4-i CD TS 4-1 rfl 4-1 CD CD g CO > g •H 0 rO 4-) m g SH !/) 0 CD (0 fa fa fa CD H rH TS rfl XI CD g co S-l -H CD > r4 4-> E H fa fa CO CO C < fa a > fa fa fa CD rH D H rH CM X) CD rfl 4-1 -H co u CO > co CO CO CTl CM ro fa LD LD CTi A O rt rt X! O CO LD O O CO LD SH CO rH fa O o o fa LO LO CM rH CO LO rt o rH CO o o ro LO o o o a rH CO o o o CD rH X! rH rfl D H LD •H — O U O O CO > LO CO rt o rH CM o cn 4J at LO rH c CD r- co CM T S rH CM LD 1 l C CD -X rH CD T S •X 00 D H O fa CO CD £ • a o S-l 0 m rH CM LD CO CM ro (0 -X rH LO r- o CD •H -X CO LO rH LD M 4-1 fa CO LD 00 LO S3 H r- r- LO T S rfl o CM o o CD fa CM o o o LO o o 0 • SH CM o o fa SH C rH CD H CD O Cd •H S3 00 cn r- 3 Z LO CO rH a co r- CM CD T S cn CM rH 4-> CD CTl o o CO > CTl o o 0 • LH g CTl o o o CD rH l 1 fa CD u rH T S CO XI CD g rfl U - H CD rH rH rH SH 4-1 E H CO ro C rt. > fa fa fa w C J a rH fa CD fa > 4-1 E H E H fa CO z <; fa H fa fa TS fa M CO CO LO CD CO 1—1 rt rH CO o LD SH -H rH OO rt rt cn rO 4-1 CD r~ rH rt o CO S3 SH D H 00 LO CTl o o C T rfl > I rt rH O CO fa E H CO r- CM o o LO rH CO S-l SH o O O 0 rH LD o OO rH rt LD rH r- — CM rH LD CM E H LD o CTl ro O rt — rH o CO o rH rH o O o A O o O o X) O LO o S-l rH o fa rH ro r~ rH LO ro o CM CM CM r~ II CO CTl 00 cn O SH LD LD rH CM fa CD rt cn cn 4-1 rH LO CM co LO cn CD H CD 1 l LD ro 4-1 o g LO LD (0 fa (0 • g SH o o •H E H (0 4-1 fa CO fa CO LO ro rt M TS M CO cn CO o CO r- CD U 0 O CO r- ro rH o 4-J CO SH CO CM CM CO CD TS SH CO LO co cn o o g C fa O r- ro ro CO rO co ro ro CM S-l 4-1 LD o o o (0 CO LO o o o fa . . CM ro CM o o o CM SH CD LD o ro r-CD 4-1 rt LD r- o 4-1 CO rH CO CM CTi CD g CO rH r- r-g -H CM CM o o rfl 4-1 CO O o o SH <n rfl W LD O o o fa CO 1 i > O fa Z fa rH rH rH rH fa O a W fa CD fa rH U CM ro XI U ro fa -H fa a > H E H E H z fa rfl < o fa > M fa fa fa A P P E N D I X 9 113 3 Q CM X l TO M m 4-1 a o> T l C CD a 0) O 4H CU SH TS CD O O M D J cu CO •H a cu 4-> CO 4- 4 o >1 5- l CO 3 CO TO -H 4-1 H TO T i rH H CM CD rH fa CD TO M r—• rH r-r - ro ON SH - H CD CM LO A LO ro CD TO 4-1 CD CO CO o X l o o rH 3 SH a co TP CO O o rH rH cr1 TO >1 o ro LO S-i CO CU EH rH o rH CM o o o rH rH ro SH M o o o fa 0 CO CD CO u LO CO TP ro CD LO rH CO CD — ON CM ON LO • • EH LO TP CO ro CTi CM CM — rH o rH o o rH rH o A o o o o X! 0 • . r - TP CO M a TP rH rH CM r- TP o CJ TP LO  "3" ro ro rH CM CD -X T i * o cc CH SH C CD M 3 2 T i CD > 0 g CD OH CD rH T i XI CD TO •H M 4-1 TO C > fa a 01 4-1 CO CD CO LO r- LO co CD TP i—I CM CO TP O O O CD CM r— h co h co r-- CD CM O O o o o E H co CD < 2 O o o •• II o SH EC CD 4-1 SH CD 0 g 4H TO SH EH (0 fa TJ S-l TO TJ c TO 4-1 CO CD 4-> TO g •H -P SH co TO fa CU S-l CD 4-> CD g TO fa Q X} TO -H S-l TO > CO CM ON CO O ON CM LO CO CO CO CM CM rH rH rO I I I LO CO LO TP o ro rH r— 00 rH LO co I-- LO TP TP CO CD TP CM ON CO CM LO CD O O O CD O O O r - o o o CM ON CO TP LO CD r- co co CO rH ON o co CD co r-CM O O rH o o o o CM O O O r— I I I cu fa CJ OH fa CD EH CO fa m HH O fa CJ X) TO -H S-l TO > 4- 1 c CD T) c CD 04 CD a SH o 4H CD SH 3 T i CD o o 5- l cu a) CO -H 04 CD -P CO 4H o IH SH TO fa A X! 0 SH CM T i CD > 0 g CD CH CD XI CD TO S-l > fa OH CD 4-1 CO o CM r -CM LO CD O CO CM O O O T i rH CD (0 SH -H o o o CT TO CO CM CD OH > I EH CO LO ON ro co rH E O TP LO O CM CD O ON CM CD I— CD ON CD CM rH rH M O o o fa 0 CO LO O CJ O CO r— TP CO o 00 ON CD — rH CO r- ON ' EH o CO CD rH rH TP LO — o o CM TP rH o o o o A o o o o XI 0 rH o o SH OH r- r- o CM — ' TP o  CJ LO CD o rH r- TP rH o ON CM CO TP II ro CO CD LO o SH o CO ro rH EC CD rH CM TP rH CM -P LO CM CM CM CD -X rH 00 CO SH CD 1 1 1 TJ -X CM CO TP O g 0 OH ro TP LO 4H TO s SH o o o EH ro CU rH CM TP r— rH TO -X rH CD o TJ SH rH CO r - rH •H •X CM rH rH SH O rH ro CD CO H-l OH CO rH rH TO S-l ON LO LO o SH TJ SH CO ON ro LO TO o O O C fa CD o 00 o Cu TO rH r - TP ON 4-1 rH o o O CO CO o o O M c rH CM ro CM o o O CD H CM 3 2 U CD O ON o ON EH CO > O, EH fa O < HH HH fa CD 4-1 4-> TO CD g g -H TO 4-1 SH co TO fa CM fa Q X) TO •H M TO > rH O LO ON CO CO TP TP O O rH ON LO CM rH rH ON O O O CO O O O rH I I I CM fa O OH fa EH 2 H Hi EH > CO EH fa < < H H O " fa CJ OH IS CM XI TO • H SH TO > -P C CD T) a CD a l CD a o 4 H | CD SH 3 TJ CD 0 0 SH cu CD CO • H 04 CD H-l CO 4H 0 >i H TO fa A XI 0 SH Cu fa a, CJ CM OH 3 col T i CD > 0 g CD 05 T i CD SH CD -P c fa OH CD 4-1 CO rH CM O O O TP O O o o r- TP LO LO o o r -0 0 TP ro TP co co co CM TP TP ON O CD LO LO ro co O CD CD CO o o LO O ro LO O CD co o o o > w H H fa 114 A P P E N D I X 9 M CO CO TS rH i—i co TS rH M ON O LO rd I—I CD rt CD rfl H CO CD (0 1—1 o CM ro •H rt rH fa rt SH • H 00 fa rH o ON SH •H o r- rt 4-1 CD r - CM A LO rO 4-1 CD rH A o LD rt rfl 4-> CD rH o LO H tt rt ON X! CM S3 H DH CO XI O ro rt SH 0H o LO ON to >i CO CO O O tr rfl rt 0 O o O cr rd >1 LO ro O DJ tH L  CD S-l • CO fa EH 00 U CO fa CO LO CO LD rH fa o rH fa O o O LO rH rH SH SH • SH u o o SH o SH O o O 0 O O CJ fa CM CJ fa LO rt O CJ ON 00 L D ON rt r- CO ON — a> rH CM CD — rH rt CM ON rt — rt rH ON O tH ON o O EH O LO EH rt o rt rt — r- o rt H LO — o CM 1-3 ro rt rt — LO o rt co ON  O <,  O fa ro O o o o A fa A CJ A o o O O O o o o o X! CD XI CD XI o rH O rH O SH XI — CD SH X) rt O co U CM rO CO CM rO DH CO LO r - fa •H — CO •H — rH CO r-SH CJ L  SH CJ rH LO CM rO • rfl • > O > ON LO CO O CD CM ON 1 O CD r- o co ON rH rt II CM O CD 4-1 II CM r- 4-1 II rt CO CM LO o SH O CO rH C O SH r- co C o SH ON o tH CM rc CD CD fa CD CD fa CD 4-1 O CD CM TS rH CM ro 44 LO CM TS rH CM O CM rt 4-1 rH r- CM CM H CD 1 SC CD -X rt SH CD 1 C CD -X rH rt CM SH CD i 1 0 g CD TS •X CO O g CD T) -X r - rt O O g MH rO a 0 fa rH 4H CO a O fa LO LO r- fa CO SH CD S-l CD s SH E-H ro a o EH CO Q o O o EH cfl CM fa fa SH SH 0 O m rH CM ro fa rH CM o CM CM TS SH rH rt ON ro * rt TJ SH ro O CO •X rH CO CO TS SH ON r - CM rH H 0 ON r - CO CD •H CO SH O o L X CD •H -X r - r - LO SH O rH r - ON LO ffl SH cn co CX) SH 4-1 fa rH rO SH ON r- SH 4-> fa LO o o rfl SH rH rH LO O T3 SH CM CM rH S3 H • TS SH ON r- S3 SH TS SH LD rt CM ON C Cd D O o TS rfl o C fa o LO TS rO o o o C W rt rH rH O rfl CD O O CD fa rfl o CD fa cfl CM O O O 4-> r— O O u 4-1 CO  0 4-1 LO O o o CO o O O 0 CO rH o 0 CO CO o o o SH SH o O O fa SH c rH o o fa SH c rH CM CO O o o o CD H CD H CD o CD O CO CO •H S3 • H SH CD CO CM rt 3 Z SH CD rt CO 3 Z SH CD CM LO r~ rt as 4-i LO D CO a CD 4-1 00 r - a CD 4-1 LO o LO o 4-1 rO ON rH CX CD TS 4-1 rfl o co CD TS 4-1 CO O  CM CM as g rH o CD 4-1 CD CD g L O rH 4-1 CD CD g LD rH o O g -H O o ro CO > g •H oo o CO > g •H LO o o O ro 4-)  o o O rO 4J r — o 0 CO 4-1 LD o o  SH Cf) o fa g SH CO • L H g SH CO . . ro fa O o o o CD rfl fa o o 0 CD CO fa rH o o O CM 1 o fa fa 1 fa fa 1 1 > CD > CD o S-l r—1 TS SH rH TS CO X) CD to X) CD g CO SH EH g rO SH -H CD CO •H CD > o fa rH rH rH S3 S-l 4-1 < fa rH rH S3 H 4-1 EH fa z fa rH rH rH rH a CO to o a CO rfl C < fa o a > fa CJ > fa fa fa fa CD fa CD fa CD fa rH fa rH W rH W X! CJ DH rH XI u DH rH CM ro XI CJ rO fa CD rfl fa EH CD rfl fa •H fa > 4-1 •H fa CO 4-1 •H H CJ > SH EH EH fa CO SH EH < CO SH EH EH z fa ro z SC fa ro z o to Z SC o fa > M fa fa > fa CJ > I—I fa fa fa A P P E N D I X 10 115 •§ • i-H VH > o '.3 rC e o o tl o o 13 o •3 o S3 cd O CO CD rH rH • 0 r— CO LO CO X l TS rO •H LO TT LO CO (0 a> O 4-1 TT OO O r - CD •H SH -H rO O CO rH CM X! (0 c rH LO LO LO O 4-1 to 0 CD CO r - CO CM O > cr c SH • . c SH rH CO Ca or O 0 O O •H ze ro CJ co O C CD •H 0 P Cn •H to O 4-1 rH TS CO CO TP CO tO S 0 X SH Cs] CO co TP rH O X! 0 to SH ro CO TP CO cu SH T3 0 os r- CO CD SH rH SH a C SH Cs] TP co LO SH O 0 DH rfl SH 0 0 0 rH 0 HH e r< -P fa 0 CO 0 0 0 O 4-1 T5 rH to C (0 X rO O •H 4-1 C c rH rH O LO to r - rH 0 rH OT TS to •H LO TP TP LO c rO rO 0) O 4-1 LO CM rH ON ro 0 H-l •H rO CO CO CD CM O TS CJ to c H r- rH r - r- c 3 0 CD CO CO r-- CM CU rO > -r—i c SH • X! CD TS rO SH 0 0 0 O EH g rH < O O O (1) O O SH Cn DH 4-1 C c 0 m cu rH C 0 u rH O CM 0 0 O SH H-l tO •H O r - CO CM 4-1 3 (0 O 4-1 CM ON CO CO OT O c o •H 4-J rO O O c 0) CD 2 H-l cu X! DH •H X CO C O •H c O c tO CJ  ON co CM CD O ON CO CO O O O O 1—I CM CO TP CD EH fa A SH CM fa Q C CD Q fa Q 3 2 X O SH DH DH rH ON CM CO O CD r- rH O O O SD o o rH r ~ O O O O TP TP O CD OO LO CO rH ON ON TP CM rH LO O [— co r— TP CM rH co ro O TP LO O O rH rH CO 00 00 O CO CO LO CO CSI H O T i CO CM CM CM 0 rH CD O CM O 0 LO O CO CM X •H rH TP CO rH •H 4-1 rH CO rH CM rH rO O r - CO r -CD OH rH CO r- ON O 0 Cs] r -•H • >—H O 0 O 0 rH CM CO TP fa A SH OH fa Q C CD Q fa Q 3 2 fa 10 > o •H 4-1 CO -H 4-1 rO 4-1 CO 0 CO 0 le 0 iab TP M TP rH LO O ON CO r0 EH r— CM ro LO ON > < TP r - to r -TP CJ ro rH CD O rH rH O tO hH 0 O O O CO LO o ro CO rH LO rH rH o rH O O tO T l ro hH CD SH 3 4-> O 3 SH 4-1 CO 10 o •H c o c CO u e o c ro U SH 00 rH CD 00 CO • H EH CM ON rH ON 0) < CD O CD TP x; CJ LO LO LO r ~ 4-1 O hH O O O 0 T) C CO to CD rH X! tO •H U tO > C o •H 4-1 CO O o CD x; 4-1 c CD CD s 4-1 CD X! CO C O •H 4-1 rO rH CD SH SH O CJ CM ro O ON CO EH ON CM ON CM < r- CM r- LO u CM O ro CO 0 . . hH O O 0 0 rH ON r - 0 rH CO CM CO CO < ON TP ro r— CJ CD CO 00 0 0 hH O 0 0 0 4-1 CO CD CC A P P E N D I X 10 116 rt CD CM rH rH CM O CM rH too rH rH CM CM CO rt r- O CTl oo ro CO < rH ON CM CO r- 00 CM rt o 00 CO CJ o rH rH o o o O rH CM o cu o rH CO fa o O O o o o O o o o X! CD 1 l i 1 1 (0 rH •H X! H ro (0 •H > ar CD > Cn 00 ro r- rH CO r- o r- rt rt CM C E H o ON rH CD rt o rH rt 00 CO (0 0 < rH 00 LO CM r- r- CD co CTi 00 fa •H CJ rH rH CM o o CM rH CM rH rH 0 4-J o X (0 fa o o o o o O o O O o CD OC 1 1 1 1 CD H fa H (0 3 o CD +J -H fa O Cn 4-1 3 rH H 0 fa C\] o LO co 00 o rH co rH rH CO 4-> fa 0 E H rH LO CM CTi rt O LO O 00 rH OT a CM LO CO O o rt LO CM 00 OO H CO CJ C\] O O rH o rH o rH O O rH o CD o rO g rH fa O o O O o O o O o O 0 X) 1 1 1 1 1 1 1 •H CD CO C fa -H 0 fa tH c CO (0 C > CJ rH rO rH rH rH ro co o CO CTl LO oo O 4-1 O E H rH CO o CO rt LO CTi LO LO 00 CD •H < O LO rH rH CTl CD rH CTl rt CM fa on OC CM ro CO CO r- LO r> r- 00 CO CO c fa O o o O o O o o o O c rO 1 1 i 1 1 1 1 1 1 O O •H fa CD CO fa rH fa CD SH TS r-l C O CO OT O EM fa £ £ £ fa H I D rH E H CM < Q Q Q £ < fa fa OT fa CJ OT fa E H fa fa CJ CD C CM OO CM LO > 0 CM oo CO rH to •H -H CO oo CM rt CD fa 4J ro oo rt rt fa Cfl SH XI fa O o o O O CD m 3 a CO fa •H g o CD •H S H 3 S H rH CO rO C J fa XI O > ro a •H a fa SH o rfl fO 0 C CM I— 00 00 > CD •H 0 CM rH rt 00 fa a •H CO rH oo rH CD E H o 4-1 00 o o O Cn c S H a rfl O O o o O CO C J a fa 0 0 SH CO X fa •H CD CO > 1 fa fa CO ft) O fa T J LO oo rH CO C •H CO CD o co rH CM < tji O S H LO LO LO O fa >i -H tO CO LO CM O XI a 3 o fa o cr o o O O c a T ) c OT fO SH CD CO 1 T S o a C J fa fi g -H 3 rO T S CD fa CD fa a fa 4-1 X w CD C rt cn LO CM rH M H > 0 OO rt r - CO tO 0 to •H •H rt CD rH 00 0 CD 4-1 4-1 rt rt LO LO •H CD fa rfl S H • C 0 X! rH 0 o o O O 0 C cfl 3 a a tO C •H g 0 rO •H H 3 n C J S H O rfl C J fa tO > > SH •H fa TS CD rfl CD fa 0 c rt LO r - r-N E H •H 0 CTl LO CM LO •H C •H rt rH LO LO T ) 0 4-1 rt O o o SH C S H tO rfl 0 O O o o T J O a C 0 ro S H fa fa CO rH CM to rt 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088530/manifest

Comment

Related Items