Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies of variation in hemlock Tsuga populations and individuals from southern British Columbia Meagher, M. D. 1976

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1976_A1 M42.pdf [ 19.76MB ]
Metadata
JSON: 831-1.0075420.json
JSON-LD: 831-1.0075420-ld.json
RDF/XML (Pretty): 831-1.0075420-rdf.xml
RDF/JSON: 831-1.0075420-rdf.json
Turtle: 831-1.0075420-turtle.txt
N-Triples: 831-1.0075420-rdf-ntriples.txt
Original Record: 831-1.0075420-source.json
Full Text
831-1.0075420-fulltext.txt
Citation
831-1.0075420.ris

Full Text

STUDIES OF VARIATION IN HEMLOCK (TSUGA) POPULATIONS AND INDIVIDUALS FROM SOUTHERN BRITISH COLUMBIA by MICHAEL D. MEAGHER B.S.F., The U n i v e r s i t y of B r i t i s h Columbia, 1957 M.Sc.F., The Un i v e r s i t y of Toronto, 1963 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF Doctor of Philosophy i n th!e Department of Forestry We accept t h i s t hesis as conforming to the required standard THE UNIVERSITY OF BRITTSH COLUMBIA A p r i l , 1976 (o) Michael D. Meagher, 1976 In p r e s e n t i n g t h i s t h e s i s in p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . It i s u n d e r s t o o d tha t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department o f /Z>m4l?z/ The U n i v e r s i t y o f B r i t i s h C o l u m b i a 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date i ABSTRACT In view of western hemlock's importance to B r i t i s h Columbia's f o r e s t economy, and the lack of knowledge of i t s v a r i a t i o n patterns, modes of i n -heritance and h e r i t a b i l i t i e s , studies were begun i n 1968 on mountain and western hemlocks c o l l e c t e d from 128 parents i n 17 locati o n s i n southern B r i t i s h Columbia. The purpose was to describe the v a r i a t i o n i n cone mor-phology, to detect hybrids and determine the species' a b i l i t y to form a r t i f i -c i a l hybrids, to screen western hemlock populations for d i f f e r e n c e s i n nursery seedling performance due to a l t i t u d e and source area within a narrow l a t i t u d i n a l b e l t , and to determine the importance of cone and po l l e n parentage to western hemlock seedling growth and c h a r a c t e r i s t i c s . Morphometric examination of bagged and open-grown western hemlock cones revealed the most stable features: scale width and distance between scale t i p and widest point. These were used to compare western and mountain hemlocks, and to t e s t for i n t r a - s p e c i f i c v a r i a t i o n i n western hemlock. Material f or the seedling population studies was reared i n two nurseries: a container nursery of the Canadian Forestry Service, V i c t o r i a , and a bare-root nursery at the Un i v e r s i t y of B r i t i s h Columbia (U.B.C). Family i d e n t i t y was maintained i n both t r i a l s . Measures of seedling dimensions and f r o s t damage were taken from the former t r i a l a f t e r one growing season, while the l a t t e r was assessed for germinative rate, "mutants", bud set, bud burst, f r o s t damage and s u r v i v a l a f t e r one and two seasons. Substantial d i f f e r e n c e s between western and mountain hemlocks were found fo r both cone and seedling features. Mountain hemlock cones are longer, are com-posed of more scales, produce rounder scales and l a r g e r , more regularly-shaped, more mucronate bracts than western hemlock. No cones i n d i c a t i n g a true hybrid were found. Seeds of mountain hemlock germinated f a s t e r , and the seedlings set and burst buds e a r l i e r , were more frost-hardy and shorter than western hemlock from comparable elevations. These d i f f e r e n c e s , plus the unsuccessful h y b r i d i s a t i o n attempt on Mount Seymour, i n d i c a t e that the species are g e n e t i c a l l y d i s t i n c t . Western hemlock population studies were conducted with open-pollinated seeds c o l l e c t e d i n f i v e areas near the 49th p a r a l l e l ; sampling i n three areas extended over the species' a l t i t u d i n a l range. Height, height:diameter r a t i o , branch number and f r o s t damage to plug-grown seedlings were inversely corre-lated with elevation, whereas maximum branch length was not. Sensitive adaptation to l o c a l conditions was i n f e r r e d from di f f e r e n c e s i n the trend of seedling mean height with elevation of source between two Coastal l o c a l i t i e s . In the bare-root nursery, seed s t r a t i f i c a t i o n accelerated germination of western hemlock, but s t i l l i t was slower than mountain hemlock. Chloro-p h y l l - d e f i c i e n t or virescent seedlings were found i n western hemlock only, but i n no pattern r e l a t e d to o r i g i n . Other "mutant" seedling types were noted. A possible adaptation of " b l u i s h " seedlings to hi g h - a l t i t u d e conditions i s discussed. Bud set was associated with l o c a t i o n and a l t i t u d e of western hemlock seed o r i g i n , but bud burst was not corr e l a t e d with e i t h e r . Strong p o s i t i v e c o r r e l a t i o n s occurred between mid-autumn bud set percent and elevation, yet differences i n percent bud set were found among populations from comparable a l t i t u d e s . Bud set was negatively, though weakly, correlated with family mean height. Frost damage i n the bare-root nursery was correlated with that found i n the plug-grown stock, was independent of family height, and was negatively c o r r e l a t e d with elevation of seed o r i g i n and mid-autumn bud set. C l i n a l adaptation with elevation was apparent f o r seedling height, bud set and f r o s t hardiness. i i i A complete d i a l l e l cross was conducted between three western hemlock trees (A, D and E) at U.B.C. to allow detection of any genetic c o n t r o l over some of the features analysed i n the population studies and of the genetic nature of the species. F i l l e d - s e e d y i e l d was reduced by s e l f i n g i n a l l cases, perhaps due to the presence of deleterious recessive a l l e l e s i n each parent. Parental combination affected germinative rate and the frequency and type of abnormal germinants, but not t o t a l germination. Quantitative and q u a l i t a t i v e parameters were derived from germinants, and from seedlings a f t e r they were transplanted to plugs i n a growth chamber. Cotyledon number and hypocotyl colour appear to be c o n t r o l l e d by many genes, whereas c h l o r o p h y l l - d e f i c i e n t types appear to be homozygous recessives i n -v o l v i n g one or two genes. Analysis of height growth beginning six weeks a f t e r transplanting indicated r e l a t i v e family height at s i x months. Estimated narrow-sense h e r i t a b i l i t y at s i x months was 0.63 for height, but only 0.01 for branch length. S e l f i n g generally reduced seedling height, except for Tree D, which displayed high general combining a b i l i t y . Genetic c o n t r o l was found also for most dimension and weight features analysed, however, evidence of maternal co n t r o l was found for some features. Pronounced environmental modification of a l l growth parameters was found. Comparing the population and controlled-cross r e s u l t s , i t i s concluded that mountain and western hemlocks are s u f f i c i e n t l y d i s t i n c t g e n e t i c a l l y that i n t e r s p e c i f i c h y b r i d i z a t i o n would be an u n l i k e l y source of genes i n a western hemlock breeding program, that western hemlock populations d i f f e r e n t i a t e r a p i d l y with l o c a l i t y and elevation, p a r t i c u l a r l y i n features regulating bud set and f r o s t hardiness, that open-pollinated f a m i l i e s represent iv the general combining a b i l i t y of the seed parent in the features assessed here, and that the heterogeneity noted between and within families i s in part genetically caused, mostly by the seed parent. The adaptive role of this heterogeneity i s discussed. The preliminary suggestions for reforestation and breeding pro-grams that can be made from these studies are that seedlings should be planted within approximately 1000 feet of the elevation of seed source, and that single-tree selection and intraspecific crossing, rather than only provenance selection, be pursued in Southern British Columbia. More detailed and long-term studies are required to determine the validity of these recommendations. V TABLE OF CONTENTS Page Chapter 1 Introduction 1 Chapter 2 S i l v i c s and Genetics of B r i t i s h Columbia Hemlocks 7 2.1 Natural Ranges and s i l v i c s of B r i t i s h Columbia hemlocks 7 2.2 Genetics of B r i t i s h Columbia hemlocks 9 2.21 I n t e r s p e c i f i c r e l a t i o n s h i p 9 2.22 I n t r a s p e c i f i c v a r i a b i l i t y 13 Chapter 3 Population Studies of Mountain and Western Hemlocks from Southern B r i t i s h Columbia 15 3.1 F i e l d sampling 16 3.2 Climates of the areas of c o l l e c t i o n 20 3.21 Mean monthly temperature (Figure 3-2) 20 3.22 Mean monthly minimum temperature (Figure 3-2) 23 3.23 Percent of annual p r e c i p i t a t i o n as snowfall 23 3.24 Minimum f r o s t - f r e e period 24 3.25 Growing season p r e c i p i t a t i o n 28 3.3 S t a t i s t i c a l analyses 30 3.4 Cone morphology of mountain and western hemlocks 31 3.41 Introduction 31 3.42 Material and methods 33 Sample s i z e - Number of cones 33 Sample s i z e - Number of scales per cone 34 3.43 Results and discussion 37 I n t r a - s p e c i f i c comparisons 37 Intra-tree v a r i a t i o n 37 Environmental influence 38 I n t e r - s p e c i f i c comparisons - 41 3.44 Summary of cone studies 49 3.5 Nursery studies of mountain and western hemlock populations 53 3.51 Container nursery study 53 materials and methods 53 Results 57 Height 57 v i Page Height: diameter r a t i o 64 Branch measures 68 Maximum branch length 68 Number of branches 73 Frost damage 75 3.52 Bare-root nursery study 79. Materials and methods 80 Results 85 Germination 85 Western hemlock 85 Mountain hemlock 90 Mutants 92 Albinos and semi-albinos 93 "Healthy" mutants 99 Western hemlock 99 Dwarf mutants 99 "Bluish" mutants 101 "Odd" mutants 102 • Mountain hemlock 103 "Weak" seedlings 104 Mutant summary 106 Nursery s u r v i v a l 107 Bud-opening study 109 Methods 109 Results 111 Bud burst stage 113 Bud burst rate 117 Bud set studies 125 Methods 126 Results: 1970 127 Results: 1971 133 Western hemlock 133 Mountain hemlock 144 1972 Frost damage 148 Methods 148 Results 149 Influence of r e p l i c a t e s 149 Influence of provenance 149 Influence of elevation of source 151 Family influence 152 Relation to height growth 153 Relation to bud set 158 Nursery vs. plug f r o s t damage 161 Summary of nursery studies 168 Germination 168 Mutants 168 Survival ••• 169 Weak seedlings 170 Seed weight 170 Height ' 171 Height: diameter r a t i o 172 Branch c h a r a c t e r i s t i c s 172 v i i Page Bud burst 172 Bud set 173 Frost damage 174 Environmental v a r i a b i l i t y 175 3.6 Inter- and i n t r a - s p e c i f i c inferences from population studies 176 3.61 I n t e r - s p e c i f i c comparisons 176 3.62 I n t r a - s p e c i f i c inferences f o r western hemlock 177 Inter-population v a r i a t i o n patterns 177 Intra-population v a r i a t i o n 182 Chapter 4 I n t e r - S p e c i f i c and I n t r a - S p e c i f i c Crosses with Western and Mountain Hemlocks 187 4.1 I n t e r s p e c i f i c crossing t r i a l with western and mountain hemlocks 188 4.11 Purpose and methods 188 Parent trees 188 Pollen c o l l e c t i o n and extraction 190 P o l l i n a t i o n 191 4.12 Results 193 Phenology 193 Cone harvest 194 Seed y i e l d 194 4.13 Discussion and conclusions 197 4.2 I n t r a - s p e c i f i c crossing t r i a l with western hemlock 199 4.21 Purpose and methods 199 Parent trees 199 I s o l a t i o n technique 200 Pollen c o l l e c t i o n and extraction 200 P o l l i n a t i o n 202 Cone c o l l e c t i o n 203 Seed ex t r a c t i o n 204 Seed X-raying 204 Seed incubation 204 Transplanting 206 Measurements 207 4.22 Results and discussion 209 Receptive date 209 Cone maturity and harvest 209 Bag y i e l d 209 Cone y i e l d 210 I s o l a t i o n e f f i c a c y 210 Seed y i e l d 211 Cone s i z e . 214 v i i i Page Number of scales per cone 215 Seed weight 218 Discussion, seed weight 220 Seed germination 221 To t a l germination 221 Germinative rate 222 a) Germinative energy 222 b) Days to reach 80% of f i n a l germination 224 c) Germination Value 225 Abnormal germinants 229 a) Inverted embryos 230 b) Stunted and weak r a d i c l e s 230 . c) Forked r a d i c l e s 231 d) Cotyledon-less seedlings 232 Seedlings per cone 234 Lethal equivalents 236 Cotyledon number 237 Cotyledon length, hypocotyl length, 100-seed weight 238 Hypocotyl colour 239 "Mutant" seeding types 242 Seedling growth 249 Height 249 Relative growth rate 249 T o t a l height 253 Correlations with 6-month heights 256 Height h e r i t a b i l i t y 260 Diameter 266 Height: diameter r e l a t i o n s h i p 267 Branch length 268 Branch number 271 Dry weights 275 T o t a l dry weight 275 Top weight 286 Leaf weight 288 Root weight 294 I n t e r r e l a t i o n s h i p s between c o n t r o l l e d -cross seedling measures 298 Individual measures 298 Ratios 301 Environmental influence 304 S u i t a b i l i t y of approach 306 Summary of seedling measure i n t e r r e l a t i o n -ships from c o n t r o l l e d crosses 307 Germinative c h a r a c t e r i s t i c s . 307 Dimensions 308 Weights 309 i x Relationships between controlled-cross and population progeny r e s u l t s Mountain hemlock - western hemlock r e l a t i o n s h i p s Western hemlock r e l a t i o n s h i p s Seed weight Germinative behaviour Mutants Height and diameter Branch measures Environmental v a r i a b i l i t y Summary of c o n t r o l l e d - cross seedling studies and genetic analysis of western hemlock Chapter 5 Summary of A l l Studies and Conclusions L i t e r a t u r e Cited L i s t of Appendices X LIST OF TABLES Table No. T i t l e Page 2- 1 Morphological comparison of western hemlock (Tsuga heterophylla (Ref.) Sarg.), mountain hemlock (?. mertensiana (Bong.) Carr.) and Tsuga x J e f f r e y i (Henry) Henry. Sources: Dallimore and Jackson (1948), Den Ouden and Boom (1965), Rehder (1940) 10 3- 1 Summary of 1968 open-pollinated seed c o l l e c t i o n s of western and mountain hemlocks by l o c a t i o n and elevation of o r i g i n 19 3-2 Trends of percent of annual p r e c i p i t a t i o n f a l l i n g as snow for Coastal areas 23 3-3 Comparison of trends of minimum f r o s t - f r e e period for the Nelson area, days annually 24 3-4 Estimated length of minimum f r o s t - f r e e period by area and elevation of c o l l e c t i o n . Approximated from figures of Baker (1944) 27 3-5 Mean growing-season p r e c i p i t a t i o n f o r areas of c o l l e c t i o n 29 3-6 Intra-tree v a r i a b i l i t y i n cone values. Linear regression values, branch t i p to base. Tree E, Totem Park. Sample 40 38 3-7 Comparison of western hemlock cone scale measures from Coastal and I n t e r i o r o r i g i n s - mean values i n mm 40 3-8 Comparison of regressions of the number of cone scales with cone length f or western and mountain hemlocks 42 3-9 Comparison of selected cone and scale values for trees 99 and 181 to mean values f or western and mountain 2 hemlocks (values, apart from Cone L., xn mm or mm 45 3-10 Dimension summary of plug-grown seedlings from 1968 population c o l l e c t i o n s 58 3-11 Regressions of seedling height (cm.) on elevation of seed source (thousands of feet) 59 X I Table No. T i t l e Page 3-12 Actual and adjusted means of longest branch i n cm. a f t e r covariance removal of mean height differences. Western hemlock only. 70 3-13 Linear regressions of adjusted branch length with change i n elevation for Nelson, Haney and Seymour col l e c t i o n s . Western hemlock only 71 3-14 Linear regressions of adjusted number of branches of western hemlock families with elevation of c o l l e c t i o n : Nelson, Haney and Seymour coll e c t i o n s 73 3-15 Percentage of plug-grown seedlings showing fros t damage by area and elevation of o r i g i n 76 3-16 Values of nursery germination for s t r a t i f i e d and u n s t r a t i f i e d western hemlock seed 85 3-17 Values of nursery germination for s t r a t i f i e d and u n s t r a t i f i e d mountain hemlock seed 91 3-18 Frequency and type of f u l l y or p a r t l y chlorophyll-d e f i c i e n t seedlings i n the nursery study by o r i g i n and parent tree 94 3-19 Frequency and percentage of chlorophyll-deficient mutant seedlings i n hemlock populations studied i n the nursery, 1970 95 3-20 Frequency and percentage of "healthy" mutant seedlings, by parental population 10° 3-21 Means of weak and unringed western hemlock seedlings, plus correlation c o e f f i c i e n t , by area of c o l l e c t i o n 104 3-22 Days between commencement and completion of bud burst by parent tree i n separate provenances from Mount Seymour 122 3-23 Mean percentage bud set by area and by elevation of parent tree by inspection date for autumn, 1970 128 3-24 Linear regressions of arcsin bud set with elevation of seed source for Nelson, Haney and Seymour. Data of October 22 and December 12, 1970 130 X l l Table No. T i t l e Page; 3-25 Mean foud-^set percentage of western hemlock provenances by inspection date, 1971 and 1972 134 3-26 Linear regressions of a r c s i n of percentage bud set with elevation of seed source f o r Nelson, Haney and Seymour. Data from October 5 and 20, 1971. Western hemlock only ^35 3-27 Linear regressions of the form: Plug seedling mean height (cm.) = a + b (Bud set Oct. 20, 1971) by area of seed o r i g i n , western hemlock f a m i l i e s only 142 3-28 Percentage of mountain hemlock seedlings with set and re-opened buds by inspection date and parent tree 145 3-29 Linear regressions of 1972 f r o s t damage r a t i n g with elevation f o r progeny of Seymour, Haney and Nelson areas. Replicates 2,3 and 4 only 151 3-30 Or i g i n and family numbers of progeny showing no f r o s t damage i n the nursery, February, 1972 157 3-31 Regressions of the form: Frost damage = a + b a (Bud set Oct. 20, 1971) for areas of western hemlock seed sources: A - Reps 2 - 4 B - Rep 1 160 3-32 Regression comparisons of f r o s t damage i n plug-grown stock and i n the nursery. Form: plug f r o s t = a + b (Nursery f r o s t ) 161 3- 33 Summary of seedling studies on open-pollinated western hemlock: resume" of s t a t i s t i c a l t e s t s A. Analysis of variance 178-179 B. Linear regressions 4T1 Mountain hemlock - western hemlock h y b r i d i z a t i o n t r i a l . Number of s t r o b i l i by p o l l i n a t i o n method 192 4- 2 Mountain hemlock - western hemlock h y b r i d i z a t i o n t r i a l . Number of s t r o b i l i i s o l a t e d and cones harvested by cross 194 4-3 Dimensions of western hemlock trees used i n 1970 crossing t r i a l 200 x i i i Table No. T i t l e Page 4-4 S e l f - p o l l i n a t e d seed y i e l d by p o l l i n a t i o n method. Trees A and E 212 4-5 Y i e l d of f u l l , empty and aborted seeds following s e l f i n g of tree A 214 4-6 Mean number of scales for bagged and unbagged cones 215 4-7 Mean fresh * weight per 100 seeds (grams) by cone and p o l l e n parent 219 4-8 Germination percentage at 50 days by cone and p o l l e n parent 221 4-9 Mean germinative energy by cone and p o l l e n parent (days) 222 4-10 Days to reach 80% of f i n a l germination by cone and p o l l e n parent 224 4-11 Germination Value by cone and p o l l e n parent 225 4-12 Average seedling y i e l d per cone following c o n t r o l l e d and open p o l l i n a t i o n , dry p o l l i n a t i o n only. (Weighted means) 235 4-13 Lethal equivalents per zygote for cone parents A, D and E, 1970 ' 236 4-14 Cotyledon numbers of progeny by parent. Mean and standard deviation (S.D.) 237 4-15 Frequency of hypocotyl colour by colour c l a s s and cross 240 4-16 T a l l y of "Mutant" seedlings i d e n t i f i e d i n growth chamber material to Sept. 30th, 1971 242 4-17 Estimates of natural s e l f - p o l l i n a t i o n i n open-p o l l i n a t e d progeny of Totem Park trees 246 4-18 C o r r e l a t i o n c o e f f i c i e n t between 6-month height and e a r l i e r height measurements by parental combination 256 x i v Table No. T i t l e Page 4-19 The number and percentage* of "bound" cotyledons by cross at 12 and 24 weeks 259 4-20 Estimates of h e r i t a b i l i t y of height growth to 6 months using data of r e p l i c a t e s selected randomly from trees A and E 263 4-21 Mean values of diameter i n mm by cone and p o l l e n parent 266 4-22 The r a t i o of longest branch to e p i c o t y l length by cone and p o l l e n parents (mm per cm) 268 4-23 Values of adjusted ^ longest branch (cm.) by cone and p o l l e n parent for c o n t r o l l e d crosses only 269 4-24 Estimates of h e r i t a b i l i t y of longest branch at 6 months following covariance removal of diff e r e n c e s i n height growth. Data f o r cone parents A and E from randomly-chosen r e p l i c a t e s 270 4-25 Mean branch number per seedling by parental combination 271 4-26 Mean branch number per seedling by parental combination following covariance removal of height di f f e r e n c e s 27 2 4-27 T o t a l dry weights at 6 months: mean (x) and standard error (S E) by cone and p o l l e n parent (gm.) 275 4-28 Values and significance''" of c o r r e l a t i o n c o e f f i c i e n t s f o r seedling dimensions and 6-month weight 281 4-29 Regression summary of t o t a l tree weight on tree dimensions by cone and p o l l e n parent - s i g n i f i c a n c e of p a r t i a l F's by v a r i a b l e 283 4-30 Regression equation of top dry weight on diameter by parental combination 287 4-31 Mean dry weight of leaves per seedling (gm.) a f t e r covariance removal of height and branch length d i f f e r -ences- 288 4-32 Comparisons of the regressions of l e a f dry weight on height (ep i c o t y l length) at 6 months for crosses i n - -volving tree E 290 X V Table No. T i t l e Page 4-33 Mean dry l e a f weight per seedling i n grams following covariance removal of diameter differences 291 4-34 Top wood - l e a f weight (gm) by parental combination 293 4-35 Mean root weight per cross following removal of diameter differences (gm.) 295 4-36 Mean and standard deviation (S.D.) of shoot : root weight (gm.) by cone and p o l l e n parent 302 xv i LIST OF FIGURES Figure No. T i t l e Page 2- 1 The natural ranges of (A) western hemlock and (B) mountain hemlock. Source: Fowells, 1965 8 3- 1 Location of western and mountain hemlock c o l l e c t i o n s f o r population studies YI 3-2 Climates of areas of c o l l e c t i o n : mean monthly and mean minimum temperatures 21 3-3 Trends of f r o s t - f r e e period and cumulative growing degree days above 42 F for the Nelson area. Source: Canada Land Inventory 25 3-4 Trend of standard deviation of cone length and width of western hemlock, and de r i v a t i o n of minimum sample s i z e . 35 3-5 Scales of tree 105, mountain hemlock, mounted for -measuring 35 3-6 I l l u s t r a t i o n of basic cone scale measurements 36 3-7 Trend of scale W 7 L from base of t i p of cone, western hemlock 36 3-8 Scale W : L r e l a t i o n s h i p for i n d i v i d u a l mountain hemlock and western hemlock cone parents, 43 3-9 Scale L^ : L r e l a t i o n s h i p f o r i n d i v i d u a l mountain hemlock and western hemlock cone parents 43 3-10 Trend of scale number with cone length f or western and mountain hemlocks 47 3-11 Comparison of " t y p i c a l " scales of western and mountain hemlock cones._ Values i n mm. 47 3-12 Relationship between mean seed weight and mean seedling height f o r 23 parent trees 61 3-13 Plan of beds i n nursery study, U.B.C. Forestry nursery 82 3-14 Comparison of mean t o t a l germination f o r s t r a t i f i e d and u n s t r a t i f i e d western hemlock seed 87 x v i i Figure. No. T i t l e . Page 3-15 Course of germination f o r s t r a t i f i e d and u n s t r a t i f i e d mountain hemlock seed, and d e r i v a t i o n of germinative energy and f r a c t i o n of f i n a l . Tree 105, Replicate 3 88 3-16 1971 bud burst - comparison of western hemlock proven-ances from s i m i l a r a l t i t u d e s 114 o 3-17 Traces of grass minimum temperature F for period Sept. 1 - Dec. 31, 1970 and 1971 !38 3-18 Association between mean height of plug-grown seedlings and percentage bud set i n nursery to October 20, 1971 141 3-19 Patterns of nursery f r o s t damage with seed source and elevation. Replicates, 2 3, 4 only 150 3-20 Relationship between mean height of plug-grown seedlings and nursery f r o s t damage for western hemlock f a m i l i e s 154 3-21 Comparison of weighted f r o s t damage ratings of Replicate 1 vs. Replicates 2 - 4 156 3-22 Association between bud set at Oct. 20, 1971 and f r o s t damage r a t i n g , Reps. 2 - 4 159 3- 23 Relationship between f r o s t damage percent of plug-grown seedlings and f r o s t damage r a t i n g of nursery-grown western hemlock seedlings, Reps. 2 - 4 162 4- 1 Location of mountain hemlock trees "A", "B", and "C" at Mount Seymour s k i h i l l used i n the h y b r i d i z a t i o n attempt with western hemlock 189 4-2 X-ray photo of unpollinated mountain hemlock seed. Tree B 196 4-3 X-ray photo of c r o s s - p o l l i n a t e d mountain hemlock seed. Tree C x western hemlock D 196 4-4 X-ray photo of wind-pollinated seed of mountain hemlock. Tree A 197 4-5 Locations of western hemlock trees used i n complete d i a l l e l crossing program, Totem Park, U.B.C. 201 4-6 Representative curves of cumulative germination of E x D seeds, showing c a l c u l a t i o n of germinative energy 223 x v i i i Figure No. T i t l e Page 4-7 Comparison of 6-month height of mutant (foreground) and normal (background) s e l f e d seedlings from cone parent "A", Totem Park 2 4 3 4-8 Comparison of mutant ( l e f t ) and normal s e l f e d seedlings from cone parent "E", Totem Park 244 4-9 Trend of r e l a t i v e growth rate of height for progeny from trees A, D and E 250 4-10 Mean 6-month heights of controlled-cross seedlings by cone and p o l l e n parent 254 4-11 The r e l a t i o n s h i p between seed weight and 6-month height for c o n t r o l l e d crosses 261 4-12 T o t a l seedling dry weight a t 6 months by cone and p o l l e n parents 276 '4-13 Relationship between seed weight and 6-month weight fo r c o n t r o l l e d cross and open-pollinated f a m i l i e s 278 4-14 The r e l a t i o n s h i p between germinative energy and 6-month seedling weight for controlled-cross and open-p o l l i n a t e d f a m i l i e s 280 4-15 The e f f e c t of environmental v a r i a b i l i t y on plant dimensions: Branch length 7 height. Tree D 305 4-16 The e f f e c t of environmental v a r i a b i l i t y on plant mass: Leaf weight per seedling. Tree A 305 xix ACKNOWLEDGMENTS This d i s s e r t a t i o n has involved the assistance of many people over a period of seven years. Without the help of a l l of them i t would not have been possible to conduct or complete the necessary studies. My most profound and sincere thanks to my wife, B i r g i t t e , for assistance i n a l l phases of the work, and her u n f a i l i n g support and good humour i n the face of the hardships involved. Thanks to our c h i l d r e n for accepting the l o s s of time I can never repay. Thanks are extended also to Professor 0. S z i k l a i of the Faculty of Forestry, Chairman of my Committee, f o r arranging some of the seed c o l l e c t i o n s , arranging f i n a n c i a l support, and providing advice when consulted during the studies, and for reviewing the various d r a f t s , yet allowing me otherwise to pursue my own d i r e c t i o n s . Professor P.G. Haddock, Faculty of Forestry, c h e e r f u l l y undertook the task of supervising my work during Prof. S z i k l a i ' s absence. His c a r e f u l review of the d r a f t and encouragement during t h i s period, as well as h i s counsel and friendship over a period of years, are acknowledged with gratitude. Other members of my research committee: Professor V. Brink, Faculty of Ag r i c u l t u r e , Professors K. Beamish, K. Cole and W.B. Schofield, Department of Botany, gave generously of t h e i r time and experience, at a l l times, p a r t i c u l a r l y i n reviewing the various d r a f t s . Assistance with the heavy amount of s t a t i s t i c a l analysis was w i l l i n g l y provided by Professors A. Kozak, and S. Smith, Miss L. Cowdell and Mrs. K. Hejjas of the Faculty of Forestry, and by Mr. M. Kovats and a s s i s t a n t s , B.C. Forest Service. Their contributions made possible d e t a i l e d analysis of the very large amounts of data c o l l e c t e d , for which I am most g r a t e f u l . XX The assistance of the Canadian Forestry Service, V i c t o r i a , p a r t i c u l a r y Mr. J . Kinghorn, i n growing the container seedlings i s acknowledged with gratitude. Many associates offered encouragement and valuable c r i t i c i s m s on aspects of the various studies conducted. Among them are Professor J . Worrall, Dr. M. H. El-Lakany, Dr. R. H. Ho, Mr. H. Marshall, and Dr. E. Falkenhagen of the Faculty of Forestry, U.B.C., and my associates i n the B.C. Forest Service, p a r t i c u l a r l y Mr. Bruce Devitt, R.P.F., now with P a c i f i c Logging Co., Mr. J e n j i Konishi, R.P.F., and Dr. Francis Yeh. The f i n a n c i a l support of the Faculty of Forestry through graduate assistantships and through the Van Dusen and B r i t i s h Columbia Forest Products Ltd. Graduate Fellowships was invaluable to my family and me. F i n a l l y , I would l i k e to thank Mrs. J . Fitzgibbon and Ms. Margaret Speerstra for undertaking the onerous task of typing t h i s d i s s e r t a t i o n , with i t s unfamiliar terms and format, with patience and good humour. x x i Explanation of Plate 1 Morphological features of Western hemlock 1. A branch with p i s t i l l a t e s t r o b i l i , natural si z e 2. A staminate s t r o b i l u s , enlarged 3. An anther, front view, enlarged 4. An anther, side view enlarged 5. A branch with staminate s t r o b i l i , natural si z e 6. A p i s t i l l a t e s t r o b i l u s , enlarged 7. A scale of a p i s t i l l a t e s t r o b i l u s , upper side, with i t s ovules, enlarged 8. A p i s t i l l a t e s t r o b i l u s , lower side, with i t s bract, enlarged 9. A f r u i t i n g branch, natural s i z e 10. A cone-scale, lower side, with i t s bract, enlarged 11. A cone-scale, lower side, with i t s bract, enlarged 12. A cone-scale, upper side, with i t s seeds, enlarged 13. Seeds, enlarged 14. V e r t i c a l section of a seed, enlarged 15. An embryo, enlarged 16. Cross section of a l e a f , magnified f i f t e e n diameters 17. Winter branch-buds, enlarged 18. Seedling plants, natural s i z e xx i i P late 1 Morphological features of western hemlock. Source: Sargent (1900) x x i i i Explanation of Plate 2 Morphological features of mountain hemlock . 1. A branch with staminate s t r o b i l i , natural si z e .2. A staminate s t r o b i l u s , enlarged 3. An anther, f r o n t view, enlarged 4. An anther, side view, enlarged 5. A branch with p i s t i l l a t e s t r o b i l i , natural s i z e 6. A p i s t i l l a t e s t r o b i l u s , enlarged 7. A scale of a p i s t i l l a t e s t r o b i l u s , upper side, with i t s bracts and ovules, enlarged 8. A f r u i t i n g branch, natural s i z e 9. A cone from the Coeur d'Alene Mountains, Idaho, natural s i z e 10. A cone-scale, upper side, with i t s seeds, enlarged 11. A cone-scale, lower side, with i t s bract, enlarged 12. A scale of a small Coeur d'Alene cone, upper side, with i t s seeds, natural s i z e 13. A scale of a small Coeur d'Alene cone, lower side, with i t s bract, natural si z e 14. A seed, enlarged 15. V e r t i c a l section of a seed, enlarged 16. An embryo, enlarged 17. Cross section of a l e a f , magnified f i f t e e n diameters X 1 X 1 X 15 X 3 X 3 X 5 X15 X15 Plate 2 Morphological features of mountain hemlock. Source: Sargent (1900) 1 Chapter 1 INTRODUCTION Present forest practices i n B r i t i s h Columbia are the subject of much c r i t i c i s m , often deservedly, Chiefly, the manner i n which forests are harvested for wood use and regenerated, p a r t i c u l a r l y when i t involves continuous clear cutting and slash burning, i s being questioned vigorously by foresters and the public at large, who see values i n forests beyond those t r a d i t i o n a l l y considered. The increasing a l l o c a t i o n of once-forested low lands for i n d u s t r i a l and domestic use, and the a d v i s a b i l i t y of retaining forests at high a l t i t u d e for protection and amenity values have placed increasing pressure on the remaining forest land base to supply our future domestic and export wood markets. Although wood volumes per acre have been high i n some parts of the Province, p a r t i c u l a r l y on the Coast, sustaining a harvest of 3.4 b i l l i o n cubic feet annually ( B r i t i s h Columbia Forest Service, 1972) l i k e l y w i l l require i n t e n s i f i c a t i o n of harvesting and management practices on the most productive s i t e s . Practices such as quick re-planting at optimum spacing for the intended end use, suppression of competing vegetation and animals, f e r t i l i z i n g and periodic thinning to maintain optimum growth rate of the remaining stems a l l contribute to the enhancement of useable wood y i e l d per acre. Consequently, a strong trend toward prompt a r t i f i c i a l restocking of denuded lands has recently gained impetus i n B r i t i s h Columbia. Seedlings produced i n Pro v i n c i a l tree nurseries have risen from 6.56 m i l l i o n i n 1961 to 48.39 m i l l i o n i n 1972, rose to nearly 65 m i l l i o n i n 1975, toward an estimated sustained nursery production of 150 m i l l i o n seedlings 2 annually. At a mean density of 400 seedlings per acre and planting costs of approximately 60 d o l l a r s per acre, t h i s implies an annual investment of 22 m i l l i o n d o l l a r s i n seedling planting. Most seedlings produced f o r p l a n t i n g come from c o l l e c t i o n s of seed from natural stands when and where the cones occur. The uncertainty of adequate seed crops i s a major d i f f i c u l t y i n orderly restocking of denuded lands. Presently, the seed on hand f o r several areas r e q u i r i n g restocking i s c r i t i c a l l y low. In addition, while s e l e c t i o n of the seed parent f o r good stem g u a l i t y i s customary, the genetic q u a l i t i e s of the p o l l e n parent are never known i n such seed c o l l e c t i o n s . Thus, plantations may contain an unacceptably high l e v e l of poorly-formed phenotypes that should be rogued from the stand early, r e q u i r i n g expensive p r a c t i c e s such as f e l l i n g or poisoning. One way to improve the q u a l i t y of both cone and p o l l e n parents i n indigenous stands i s to e s t a b l i s h seed-production areas. These are naturally-occurring stands containing a high proportion of well-formed i n -d i v i d u a l s that have been rogued of poorly-formed types within a reasonable distance of the p l o t l i m i t s , depending on the l i k e l y p o l l e n f l i g h t by the species i n that area. C u l t u r a l p r a c t i c e s such as adding mineral f e r t i l i z e r s and removing ground vegetation u s u a l l y are employed to enhance seed y i e l d s . The s e l e c t i o n of well-formed parents from a broad region of reason-ably uniform climate and t h e i r propagation i n a p l a n t a t i o n managed f o r the express purpose of producing seed (a seed orchard) has become increas-i n g l y wide-spread i n recent years as a method of producing seed crops more re g u l a r l y than those a v a i l a b l e n a t u r a l l y . In addition, the concentration of 3 s e l e c t i o n s from a wide range of l o c a l i t i e s permits the creation of genetic combinations that could not occur i n nature because of the distance between t h e i r parents. The r e s u l t i n g seedlings might perform better than ones grown from l o c a l seed c o l l e c t i o n s because the broader genetic base of t h e i r parents w i l l make them capable of occupying more diverse habitats. The costs of e s t a b l i s h i n g and maintaining seed orchards are high, with very l i t t l e seed of Douglas-fir forthcoming during the f i r s t ten years. C a r l i s l e and Teich (197(D) estimated that only a modest increase (2% to 5%) i n merchantable timber y i e l d would more than o f f s e t the genetic research needed to produce such gains. Recently, Namkoong (1973) outlined programs of breeding Coastal Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco), two " I n t e r i o r " spruces- (Engelmann spruce', Picea  engelmannii Parry and white spruce, P_. glauca (Moench) Voss) , and lodgepole pine (Pinus contorta var. l a t i f o l i a Engelm.) that f e l l within or above these l i m i t s . These pr e d i c t i o n s are based on a knowledge of the genetic v a r i a b i l i t y and breeding c h a r a c t e r i s t i c s of the species stemming from years of research. The p r i n c i p a l species native to B r i t i s h Columbia not included i n Namkoong's summary i s western hemlock (Tsuga heterophylla (Raf.) Sarg.). Western hemlock i s one of the most important tree species i n B r i t i s h Columbia. The standing, mature, sound-wood hemlock volume estimated i n the most recent P r o v i n c i a l inventory, including mountain hemlock (T. mertensiana (Bong.) Carr.), was 62,805.2 m i l l i o n cubic feet (MM c . f . ) , representing about 32.2% of a l l sound mature wood i n the Province (B.C. Forest Service, 4 (1972). This exceeds the volume of a l l other i n d i v i d u a l species and of a l l other tree genera but spruce, which includes Engelmann, white, S i t k a (Picea s i t c h e n s i s (Bong.) Carr, and. black (P_. mariana (Mill.) B.S.P.) spruces. The volume of hemlock harvested i n 1974, the l a s t year for which f i g u r e s are a v a i l a b l e , was 470.5 MMc.f. This constituted 22.2% of the t o t a l p r o v i n c i a l harvest and was not surpassed by any other species. The prominence of hemlock i n the p r o v i n c i a l economy i s f a i r l y recent. E a r l y inventories d i d include i t , although i n i t i a l l y i t was considered a "weed" tree by some, us e f u l mainly for the tannin i n i t s bark. This was based on experience with eastern hemlock (Tsuga canadensis L.) which i s a c l o s e l y - r e l a t e d , widespread species of eastern North America with weak, brash wood. From i t s f i r s t report i n 1914 (Forest Branch Annual Report), to 1973 (Forest Service Annual Report), the P r o v i n c i a l harvest ;of western hemlock has increased by a factor of 41.4 from 65.61 MM f.b.m. to about 2,715 MM f.b.m. During the same period the t o t a l P r o v i n c i a l wood harvest has r i s e n by a factor of 18.0 over the 1914 l e v e l ( i b i d . ) . Because of the high remaining volume and i t s s u i t a b i l i t y f o r both solid-wood use and as a f i b r e i n pulps, western hemlock w i l l continue to be an important component of B r i t i s h Columbia's f o r e s t economy. Regeneration of western hemlock i s overwhelmingly by natural s e e d f a l l . Planting has been confined mainly to highly-productive s i t e s with severe competition from shrubby "bursh" species. Sowing f o r production of 2-0 nursery stock was 1.579 m i l l i o n i n 1968 and had r i s e n to only 1.845 m i l l i o n i n 1971. More seedlings would have been grown f o r r e f o r e s t a t i o n i f seedling s u r v i v a l and growth, both i n the nursery and following outplahting 5 of conventional bareroot stock, had been better. The development of "container" f a c i l i t i e s , permitting more precise c o n t r o l of growing conditions, promises to boost the production of hemlock seedlings dramatically. For example, sowings i n 1976 are intended to produce 7.0 m i l l i o n seedlings, 99% of them i n containers. Recently, plans to begin tree improvement i n t h i s species have been made i n Great B r i t a i n (Gardiner and Faulkener, 1966), the United States (Wheat, 1972) and Canada (Packee and Handley, 1972, Piesch, 1974). Tree improvement by s e l e c t i o n and breeding depends upon a knowledge of the kinds and degrees of genetic c o n t r o l over the t r a i t s desired. There are no long-term r e s u l t s from family or provenance t r i a l s i n B r i t i s h Columbia to guide us i n the improvement of western hemlock. In an attempt to provide some bases on which to plan future work on the genetic improvement of western hemlock, a number of studies of hemlock populations and i n d i v i d u a l s from southern B r i t i s h Columbia were begun i n 1968. Morphological and phenological studies were conducted to determine both i n t e r - s p e c i f i c d i f f e r e n c e s between mountain and western hemlocks and i n t r a - s p e c i f i c d i f f e r e n c e s i n western hemlock. These involved both open-p o l l i n a t e d population seed samples, and c o n t r o l - p o l l i n a t e d seeds from selected trees. The population studies were intended to determine: i) features of cone morphology and seedling performance useful i n separating western and mountain hemlocks and i d e n t i f y i n g hybrids i i ) inter-population d i f f e r e n c e s i n western hemlock seedling performance associated with l o c a t i o n and a l t i t u d e of seed source within one l a t i t u d i n a l b e l t 6 i i i ) the degree of intra-population v a r i a b i l i t y in western hemlock germination and seedling performance due to cone parent Inter- and intra-specific crosses with selected mountain and western hemlocks were conducted to determine: iv) The a b i l i t y of western and mountain hemlocks to be hybridized a r t i f i c i a l l y v) The importance of cone and pollen parentage to western hemlock seed yield,, germinative characteristics, seedling growth and morphology. 7 Chapter 2 SILVICS AND GENETICS OF BRITISH COLUMBIA HEMLOCKS 2.1 Natural ranges and s i l v i c s of B r i t i s h Columbia hemlocks The natural ranges of western and mountain hemlocks are presented i n Figure 2-1. Western hemlock extends over 23 degrees of l a t i t u d e (38-51 degrees north) and 38 degrees of longitude (113 to 151° west). I t e x h i b i t s a d i s j u n c t area i n the humid I n t e r i o r region. E a r l i e r maps (Sargent, 1900, Harlow and Harrar, 1958, Preston, 1961) showed these areas joined by a t h i n band l y i n g near the Canada-U.S. border. The range i n recent maps (Fowells, 1965, Harlow and Harrar, 1968, L i t t l e , 1971) shows a d e f i n i t e separation between the Coastal and I n t e r i o r areas, which both contain small d i s j u n c t i o n s . Within both the Coastal and I n t e r i o r portions of i t s range, western hemlock occurs from low to high elevations (Fowells, 1965). On the Coast, the upper l i m i t i s the beginning of the upper subassociation of the mountain hemlock - sub-alpine f i r (Abies lasiocarpa (Hook) Nutt.) as s o c i a t i o n , which l i e s between 3000 and 4000 fee t on the lower Coast (Peterson, 1964). Western hemlock i s associated with mountain hemlock throughout most of i t s range (Figure 2-1). The l a t t e r species occurs at generally higher elevations than the former, but t h e i r d i s t r i b u t i o n s overlap somewhat i n the upper l e v e l s of western hemlock's d i s t r i b u t i o n . Above t h i s , mountain hemlock occurs scattered i n d i v i d u a l l y or i n clumps throughout a sub-alpine ass o c i a t i o n . Western hemlock's shade tolerance i s high, but i t i s very s e n s i t i v e to droughty s o i l s , and i s more s e n s i t i v e to f r o s t than Douglas-fir (Duffield, Figure 2-1. The natural ranges of (A) western hemlock and (B) mountain hemlock. Source: Fowells, 1965. 9 1 9 5 6 ) . Even so, the absolute minimum temperatures encountered within o o i t s range l i e between r-47 C i n the I n t e r i o r and - 2 5 C on the Coast of B.C. Absolute maximum summer temperatures can reach 38°C. Provided the moisture supply to the roots i s not interrupted and i s not too r i c h i n mineral nutrients, western hemlock can withstand such high -temperatures for short periods (Krajina, 1 9 6 9 / 7 0 ) . I t germinates r e a d i l y on a v a r i e t y of sub-st r a t e s , but survives best on acid humus, perhaps dying e a r l y on "moder humus i n r i c h s o i l s " (ibid.) Mountain hemlock's d i s t r i b u t i o n i s c l o s e l y r e l a t e d to areas that become covered i n the early autumn by heavy, wet snow, which may remain u n t i l l a t e J u l y ( i b i d . ) . Krajina ( 1 9 6 9 / 7 0 ) considers the two species to function very s i m i l a r l y : both are shade-tolerant, moisture-demanding species able to function on r e l a t i v e l y low l e v e l s of nut r i e n t s . 2 . 2 Genetics of B r i t i s h Columbia hemlocks 2 . 2 1 " I n t e r s p e c i f i c r e l a t i o n s h i p Some morphological features of western and mountain hemlock are presented i n Plates 1 and 2 . More features and those of t h e i r putative hybrid are compared i n Table 2 - 1 . Differences between them e x i s t i n the following features: l e a f d i s p o s i t i o n , c r o s s - s e c t i o n a l shape, stomatal d i s t r i b u t i o n and markings, l e a f margin, winter twig colour, the shape of the basal bud scales, general proximity of male to female s t r o b i l i , habit of immature female s t r o b i l i (erect or pendent), s i z e and shape of mature female s t r o b i l i , seed s i z e , colour and markings. P i l g e r ( 1 9 2 6 ) assigned them to d i f f e r e n t sections of Tsuga: T_. heterophyl l a i n Eutsuga, and T_. mertensiana i n Hesperopeuce. Table 2-1 Morphological comparison of western hemlock (Tsuga heterophylla (Raf.) Sarg.), mountain hemlock (T. mertensiana (Bong.) Carr.) and Tsuga X J e f f r e y i (Henry) Henry. Sources: Dallimore and Jackson (1948), Den Ouden and Boom (1965), Rehder, (1940) Feature W e s t e r n h e m l o c k M o u n t a i n h e m l o c k Tsuga X J e f f r e y i Leaves D i s p o s i t i o n Dimensions Cross-s e c t i o n a l Markings Mainly 2-ranked i n shade also A l l leaves generally upswept, radi a t e from branches i n l i g h t ; strong tendency for "rosettes" i n "rosette" whorls on o l d of leaves to form on short trees or short upper shoots shoots, regardless of expos-where f u l l y - exposed. Leaves ure. Leaves usually curved Margins u s u a l l y s t r a i g h t Upper leaves on twig much shorter than lower ones (6 vs. 15 mm) x 1.5-3 mm Flattened, grooved above and ridged beneath Two broad bands of stoma-t a l rows below, c l e a r margin. Colour dark green above, near white below Serrulate forward. Upper leaves nearly equal to lower, 10 - 25 mm x 2 -3.5 mm Less f l a t t e n e d , diamond-shaped or semi-circular. Upper surface ridged or grooved Stomatal rows on a l l sur-faces, stomatal bands i n -conspicuous, margins le s s d i s t i n c t . Colour glauc-ous green to grayish E n t i r e At r i g h t angles to shoot. Not curved, but poin t forward About 12 mm x 2 mm Upper surface f l a t -tened s l i g h t l y to near the apex F u l l stomatal bands below, a few broken l i n e s near the apex above. Colour dark or d u l l green. Tip h a l f obscurely toothed Winter twigs Generally slender, yellow-brown to brownish or red-brown. Pubescent 4 to 6 years, long & short hairs intermixed Stouter, red-brown to purp-l i s h brown. Pubescent 2 -3 years Table 2-1 Feature (continued) W e s t e r n h e m l o c k Buds Small (1-2 mm), blunt yellow-i s h to red-brown; basal scales short and blunt, reach-ing about h of bud length Bark Scaly ridges when mature, scales brown-gray with cinna-mon-purple inner face S t r o b i l i immature Staminate P i s t i l l a t e -mature ( P i s t i l l a t e ) Seeds Size - seed wing Colour Markings Dark reddish, turning yellow as p o l l e n matures, stalked. Borne i n s o l i t a r y or c l u s -tered a x i l l a r y buds, often close to female buds Terminal buds on short shoots, drooping, r a r e l y erect. Bracts rounded, shorter than scales. Scales red to red-purple, l a t e r green. 3.5-5 mm long Woody ovoid cone 15-25 mm x 6 to 10 mm. Scale shape spatu-l a t e , width about h length. Margin i r r e g u l a r to erose, pub-erulous near outer t i p . Bract lozenge-shaped, t i p blunt, colour dark brown 2 -3.5 mm long 2 to 3 times seed •. length Medium brown to reddish brown Resin v e s i c l e s occasional M o u n t a i n h e m l o c k Tsuga X J e f f r e y i Larger (2—3 mm), more pointed red-brown to purplish-brown. Basal scales larger and sharper, reaching about 3/4 of bud length. Tips spiny Scaly ridges when mature, scales grey-brown with more p u r p l i s h inner face Dark red-purple, becoming y e l l -owish as p o l l e n matures, stalked. Buds s o l i t a r y or clustered. More dist a n t from most females Terminal buds, erect during p o l l -i n a t i o n , l a t e r pendant. Bracts with mucronate t i p , larger than scales. Scales deep purple or red-purple. Retaining colour. 7 to 11 mm long 20-75 mm x 13-16 mm woody c y l - Woody cone c y l i n d r i c i n d r i c a l cone. Scale shape round- to oblong, to 50 mm i s h to obovate, width about equal to length. Margin smooth to i r r e -gular, noticeably puberulous near t i p . Bract more shouldered, t i p sharp, colour dark brown or purple 2.5-4 mm long 8-11 mm long by 2. 5-4*-mm.. Dark brown 1 to 2 large r e s i n v e s i c l e s 12 A putative natural hybrid between the two species was c o l l e c t e d i n B r i t i s h Columbia over a centry ago. I t was c a l l e d J e f f r e y i (Henry) Henry (Henry and Flood, 1919). Later Rehder (1940) l i s t e d i t as a hybrid, a p r a c t i c e followed since. I t s features appear intermediate to those of the parental species (Table 2-1). The status of mountain hemlock was appraised through a se r i e s of inv e s t i g a t i o n s on herbarium and arboretum materials conducted i n France (Van Campro-Duplan and Gaussen, 1948, Vabre, 1954, Vabre-Durrieu, 1954, Van Campo, 1955). They considered the following features: general morphology, karyotype, p o l l e n grains, and morphological features of the seedling progeny. Van Campo-Duplan and Gaussen (1948) considered mountain hemlock an intergeneric hybrid i n v o l v i n g western hemlock and S i t k a spruce (Picea s i t c h e n s i s (Bong.) Carr.), naming i t Tsugo-Picea Hookeriana Murr. T. J e f f r e y i was re-named Tsugo-Piceo-Tsuga J e f f r e y i (Henry) nob. Tsuga c r a s s i - f o l i a Flous was considered a hybrid of T. mertensiana and Picea engelmanii, and Tsuga longibracteata Cheng, was considered to be another intergeneric hybrid i n v o l v i n g Tsuga chinensis (Franch) P r i t z e l and Ke t e l e e r i a Evelyniana Masters. I f t h e i r a nalysis of the o r i g i n of t h e " a t y p i c a l " hemlocks forming the Hesperopeuce section of the genus i s correct, they are the r e s u l t s of a remarkable p r o c l i v i t y for some hemlocks to form interpgeneric hybrids. Among c o n i f e r s , only Cupressocypar.is Dallimore has been accepted recently as an intergeneric hybrid i n v o l v i n g Chamaecyparis nootkatensis (D. Don) Spach. and various Cupressus species ( M i t c h e l l , 1972). These hybrids arose and were : . found under favourable c u l t u r a l conditions. The presumed hybrid between western and mountain hemlocks has been ~ •.•:.„. found i n c u l t i v a t i o n , but f a i r l y extensive, although 13 intermittent, searching of areas where the species occur n a t u r a l l y has produced only vague reports of i t s natural occurrence. Recently, Taylor (1972) described the morphological and phytochemical c h a r a c t e r i s t i c s of western and mountain hemlocks and some western American spruces. Only 3 of 233 hemlock samples were intermediate to western and mountain hemlock both morphologically and chemically, although many others were morphological intermediates. Regarding the p o s s i b i l i t y of an i n t e r -generic o r i g i n to mountain hemlock, Taylor (1972) ou t l i n e d the chemical a f f i n i t i e s of the hemlocks and spruces studied. Mountain hemlock was the hemlock species most c l o s e l y r e l a t e d to the spruces. Attempts to t e s t the r e l a t i o n s h i p between western and mountain hemlocks by creating the hybrid a r t i f i c i a l l y have been made (Franklin, 1969, Taylor, 1972, Sorensen, 1974) without production of f i l l e d seed. Therefore, the p o t e n t i a l of mountain hemlock as a source of genes i n a western hemlock breeding program may be l i m i t e d , but t e s t i n g i s needed. However, there may be s u f f i c i e n t v a r i a b i l i t y i n western hemlock to permit rapid gains without h y b r i d i z i n g . 2.22 I n t r a s p e c i f i c v a r i a b i l i t y The B r i t i s h Forestry Commission (Lines and Aldhous, 1960, 1961, 1962, Lines and M i t c h e l l , 1969) has reported the performance of 18 provenances (2 from Great B r i t a i n , and therefore of unknown source i n i t i a l l y ) planted throughout B r i t a i n . The North American provenances were from both the Coastal and I n t e r i o r p o r t i o n of the species' range. Nursery observations of time of bud set, f r o s t hardiness and height growth indicated that these measures were associated l a r g e l y with l a t i t u d e of seed o r i g i n , and perhaps 14 also with a l t i t u d e for one high-altitude Coastal source (Enumclaw, 3500 feet ) . Provenances that originated i n high latitudes or altitudes ceased height growth and set buds early, producing generally shorter and more frost-hardy seedlings. An Inte r i o r provenance (Shuswap Lake) set buds r e l a t i v e l y early, giving i t a 1962 growing season of 117 days, but t h i s was longer than the growing season (104 days) of the Enumclaw provenance. After 5 to 7 years on the planting s i t e s , height values were much affected by plantation-provenance interaction. In general, best height growth was achieved by provenances orig i n a t i n g closest to the l a t i t u d e of the planting s i t e . The high-altitude Enumclaw provenance behaved l i k e one originating further north. These res u l t s indicate substantial i n t r a - s p e c i f i c v a r i a b i l i t y i n western hemlock associated with l a t i t u d e and a l t i t u d e . In view of the uncertainty of i t s relationship to mountain hemlock outlined e a r l i e r , and the lack of family information i n the B r i t i s h provenance t r i a l , the studies outlined i n Chapter 1 were conducted. Chapter 3 POPULATION STUDIES OF MOUNTAIN AND WESTERN HEMLOCKS FROM SOUTHERN BRITISH COLUMBIA Study of f o r e s t tree populations has been underway more than one hundred years (Langlet, 1962). In 1823, de Vilmorin found evidence of differences i n performance of tree seedlings r e l a t e d to t h e i r geo-graphical o r i g i n (Langlet, 1971b). Subsequent experience with many tree species i n many areas of the world has led to recognition of the f a c t that s t r i k i n g d i f f e r e n c e s i n adaptation to c l i m a t i c pattern have developed between populations of wide-ranging species (Callaham, 1964). The importance of knowing the r e l a t i o n s h i p s between species and the patterns of v a r i a t i o n within them before planning a breeding pro-gram are emphasized by Wright (1962). The studies described i n t h i s Chapter were conducted to explore some aspects of v a r i a t i o n i n mountain hemlock and western hemlock populations i n southern B r i t i s h Columbia. S p e c i f i c a l l y , they were intended to determine the c h a r a c t e r i s t i c s of the cones and seedling performance of each species to permit more accurate determination of t h e i r hybrids than possible to date (Table 2- 1 ), to gauge the genetic r e l a t i o n s h i p between the two species, to determine the kinds of inter-pop u l a t i o n d i f f e r e n c e s i n western hemlock progeny associated with seed source l o c a t i o n and a l t i t u d e , and the r e l a t i o n s h i p to c l i m a t i c trends, to detect the degree of intra-population v a r i a b i l i t y i n western hemlock seedling performance a t t r i b u t a b l e to cone parent, and to draw appropri-ate inferences regarding the genetic nature and a d a p t a b i l i t y of western hemlock populations from d i f f e r e n t a l t i t u d e s i n a narrow l a t i t u d i n a l b e l t 16 3.1 F i e l d sampling Current-season cones were c o l l e c t e d from stands i n separate areas and elevations i n the autumn of 1968 to provide cone and seed, material f o r studies of cone morphology and i n t e r s p e c i f i c , interpopulation and intrapopulation seedling performance. The main areas of c o l l e c t i o n l i e p a r a l l e l to the 49th p a r a l l e l of north l a t i t u d e i n B r i t i s h Columbia, and are: Nelson, longitude 117°20', Coast-Interior t r a n s i t i o n (Hope slide-Manning Park),, longitudes 120° -121°20', Haney and Mount Seymour i n the Lower Mainland, longitudes 122° 30" and 123°00', and Vancouver Island (Caycuse and Paterson Lake), longitudes 124°25', to 125°30' (Figure 3 - 1 ) . Within the Nelson, Haney and Mount Seymour areas, c o l l e c t i o n s were made from stands at separate elevations. Before cones were c o l l e c t e d , the v i c i n i t y was checked to locate stands at approximately 500-foot i n t e r v a l s from which f i v e or more trees could provide cones. Lines (1967) recommended that the number of parent trees per l o c a t i o n f o r provenance t e s t s be determined by the v a r i a b i l i t y of the species i n the area, with a "barely adequate" minimum of 10 trees per stand, and a "better number" of 25 to 50. Such trees should be selected at "...an i n t e r v a l that exceeds the normal s e e d - f a l l distance". Lacking knowledge of the natural v a r i a b i l i t y of western hemlock i n any area, and lacking the resources to handle the material from the number of trees per l o c a t i o n indicated by Lines (1967), the minimum number of trees per l o c a t i o n was set a r b i t r a r i l y at f i v e . More were c o l l e c t e d i f time and cone crop permitted. Figure 3-1. Location of western and mountain hemlock collections for population studies. Cones were collected from the upper crown, i n most cases, and kept separate by parent tree. The distance between the trees was variable, and determined only by the size of the cone crop. Adjacent trees often were sampled, increasing the chance that they might have originated from the same seed parent. Thus, the relationship between the progeny of trees i n some col l e c t i o n s could be greater than half s i b l i n g s be-cause of the increased chance of p o l l i n a t i o n s between closely-related trees. Namkoong (1966) has broken the kinds of bias i n r e s u l t s from such collections into four: (i) the f r u i t f u l parents only can be sampled, l i m i t i n g the breadth of material to a portion of that available, ( i i ) the parents may be related, producing inbred seedlings (as noted above), ( i i i ) males (pollen sources) may be related, producing progeny from one female that are closer related than half sibs, (iv) the average relationship between seedlings from one cone parent w i l l be biased toward the male producing the most successful f e r t i l i z a t i o n s . Since the purpose of t h i s study was p a r t l y to assess the importance of elevation of c o l l e c t i o n of open-pollinated seed, such as are made generally i n forest practice, on the early performance of t h e i r progeny, point (i) i s not considered a serious deficiency of t h i s study. The other biases must be accepted because lack of time to investigate the areas more thoroughly precluded c o l l e c t i n g from trees separated more widely within a stand. Cones collected from each tree were placed i n bags and stored under shelter outdoors u n t i l removed for seed extraction. Collecting began i n l a t e August at Nelson, and ended on Mount Seymour i n early December. One hundred and twenty-eight trees were sampled (Table 3-1 ). 19 Table 3- Summary of 1968 open-pollinated seed c o l l e c t i o n s of western and mountain hemlocks by l o c a t i o n and e l e -vation of o r i g i n . L o c a l i t y Total Latitude Longitude Elevation (ft.) trees Species"1 Nelson-6-mile Nelson-Apex Nelson-Mine Haney-232 St. Haney-Arboretum 49° 30' 49° 30' 49° 30' 49° 20' 49° 20' 117° 20' 117° 20' 117° 20' 122° 35' 122° 35' 2000 3500 5200 100 500 5 5 5 12 12 W.H. W.H. W.H. W.H. W.H. Haney-Roads E S S Haney-Rock P i t Seymour-Dollarton 49° 20' 49° 20' 49° 21' 122° 35' 122° 35' 123 1300 1800 50 12 W.H. 11 W.H. 9 W.H. Seymour-St. Pius Church 49° 21' 123 400 W.H. Seymour-Power Line Mi.lh 49° 21' 123 1000 W.H. Seymour Mi.3^ Seymour-T. V. Towers 49° 21' 49° 21' 123 123 1900 2750 6 7 W.H. W.H. Seymour-Ski 49° 21' 123 3200-3300 M.H. Caycuse 48° 50' Paterson Lake 50° 10' Hope S l i d e Manning Park 1 49° 15' 49° 05' Species abbreviations: 124° 25' 125° 30' 121 15'-20' 120 05' 600 800 3500 4100 "W.H." = Western hemlock "M.H." = Mountain hemlock 11 5 7 2 W.H. W.H. W.H. & M.H. W.H. Seed extraction was done by hand a f t e r cones opened indoors. Cones were retained separately by parent tree and stored. The cleaned seeds were returned to storage at 1°C u n t i l needed for further t e s t s . 3.2 Climates of the areas of c o l l e c t i o n I t was d i f f i c u l t to obtain r e l i a b l e c l i m a t i c data for many of the elevations sampled because few weather stations are located as high as the c o l l e c t i o n s made above the v a l l e y f l o o r s . The Mount Seymour and West Kootenay (Nelson) areas are exceptions, with a s t a t i o n on Mount Seymour at 2700 feet, and stations above 6000 feet near Nelson. Values of temperature, p r e c i p i t a t i o n and percent of annual pre-c i p i t a t i o n f a l l i n g as snow were obtained f o r the c l o s e s t appropriate s t a t i o n from "Climate of B r i t i s h Columbia, Report for 1966" of the Department of A g r i c u l t u r e , Province of B r i t i s h Columbia (Anon., 1968). The report of that year was used because i t i s the l a s t of that s e r i e s to present long-term averages. A d d i t i o n a l information on temperature lapse rates, f r o s t - f r e e periods and growing degree days were supplied by the Climatology D i v i s i o n , Canada Land Inventory i n B r i t i s h Columbia. These data are presented and discussed i n an attempt to characterize the climates of the areas of c o l l e c t i o n s , but the f a l l i b i l i t y of using short-term averages, and extrapolating from c l i m a t i c data, i s r e a d i l y admitted. 3.21 Mean monthly temperature (Figure 3.-2 ) Vancouver K i t s i l a n o was used f o r the base of Mount Seymour since other stations nearby lacked complete information. The data from 21 Legend: Nelson A Mean monthly A Min. monthly mean O -Haney J F M A M • J J A S O N D J M o n t h Figure 3-2. Climates of areas of c o l l e c t i o n : mean monthly and mean minimum temperatures. 22 K i t s i l a n o show that mean monthly temperature r i s e s gradually and s t e a d i l y from March to July, that J u l y and August have the same mean values, despite the shorter day length i n August, and that temperatures change more r a p i d l y i n the autumn than they do i n the spring. Winter means (December-February) are a l l above freezing, r e f l e c t i n g the i n -fluence of the sea. The trend f o r Hollyburn ridge, elevation 3120 feet, ten miles west of Mount Seymour, i s s i m i l a r to that f o r K i t s i l a n o , although always colder. Haney's trend i s also s i m i l a r to K i t s i l a n o ' s , but s l i g h t l y cooler, even i n summer. The Nelson trend (1980 feet) i s of cooler temperatures than K i t s i l a n o during January - A p r i l and a f t e r October, but hotter during J u l y . The two stations appear to be equal during the r e s t of the year. From data analysed by the Climatology D i v i s i o n , Canada Land Inventory, the trend of mean temperature w i t h . a l t i t u d e during the growing season f o r the Nelson area i s of an increase varying between o o 4.95 F per 1000 feet i n A p r i l to 1.75 F i n June. This occurred between stations 4745 fee t d i f f e r e n t i n a l t i t u d e (Harrop, 1755 f e e t , and "Micro 7000", 6500 f e e t ) . Anomalous r e s u l t s of t h i s kind are common on mountainous areas, r e f l e c t i n g thermal b e l t s caused by a i r drainage and pooling (Schmidt 1968). The s t a t i o n c l o s e s t to Caycuse, Cowichan Lake Forestry, and that c l o s e s t to Paterson Lake, Campbell River, give traces of mean temp-erature nearly i d e n t i c a l to Haney, except for s l i g h t d i f f e r e n c e s i n January and February. The values f o r Cowichan lake should be close to those for Caycuse, which s i t s further west on the lake, but Paterson Lake i s about 12 miles west of Campbell River away from the sea and near the mountain spine of Vancouver Island. Consequently, i t may be warmer i n summer and cooler i n winter than the values show here. 3.22 Mean monthly minimum temperature (Figure 3- 2) R e l i a b l e data were a v a i l a b l e f o r only Haney and Nelson. The trends show that Nelson i s colder than the Coastal areas f o r a l l months but June to August i n c l u s i v e . This l i k e l y r e f l e c t s the higher basal e l e -vation of the Nelson s t a t i o n r e l a t i v e t o the others and the moderating influence of the ocean at the Coastal s t a t i o n s . 3.23 Percent of annual p r e c i p i t a t i o n as snowfall Data f o r t h i s parameter were supplied by the Canada Land Inventory. The best coverage was a v a i l a b l e f o r the Haney area, with poorer coverage from the Vancouver and South-east Vancouver Island areas. Linear r e -gressions of percent snowfall with elevation were ca l c u l a t e d f o r the areas separately,•and combined. They appear i n Table 3-2. Table 3-2 Trends of percent of annual p r e c i p i t a t i o n f a l l i n g as snow f o r Coastal areas Area Station Regression equation r 2 S i g n i f - ^ E l e v a t i o n Percent snow icance range (feet) Haney 1140 1.828 + 5. 07 -3 ' 10 J E l . 0.87 * * Vancouver 3100 2.434 + 2. 36 • io" 6 E l . 2 0.97 * * Vancouver Island 1260 1.235 + 5. 17 -3 • 10 J E l . 0.99 * * A l l pooled 3100 3.628 + 3. 13 • io" 6 E l . 2 0.94 * * 1 ** S t a t i s t i c a l l y s i g n i f i c a n t at the 1% p r o b a b i l i t y l e v e l A c l e a r and strong increase i n the importance of snowfall occurs i n each area, but p a r t i c u l a r l y i n the Vancouver region, where the trend with elevation i s not a l i n e a r one, but increases exponentially. This i s s i m i l a r to the observation of Schmidt (1968, Table 3) that snow depth increased more r a p i d l y per u n i t of elevation near the top of Crest Mountain than at the base. 3.24 Minimum f r o s t - f r e e period This value was chosen f o r study because of the l i k e l i h o o d that occasional severe stress has a more profound e f f e c t on plant population than the "average" conditions expressed by c l i m a t i c mean values, which mask considerable v a r i a b i l i t y . The minimum f r o s t - f r e e period i n a l l areas decreased as elevation increased. Values read from an approximate trend drawn f o r Nelson area stations (Figure 3-3 ) were compared to those derived from Baker's (1944) curve f o r northern Idaho. The r e s u l t s are presented i n Table 3-3 Table 3- 3 Elevation (feet) Comparison of trends of minimum f r o s t - f r e e period for the Nelson area, days annually. S o u r c e Baker 1944 Canada Land Inventory 1972 Brink 1947 2000 3000 4000 5000 6000 144 120 93 66 33 124 112 100 88 76 97 (Crescent Valley) 144 (Nelson) The two trends agree f a i r l y w e ll only f o r elevations of 3000 and 4000 feet. The C.L.I, data f a l l below Baker's estimate at 3000 fee t and above 25 Legend + Fr o s t - f r e e period Growing degree days > o ,Q as cn >i rd ti cu CD U tr> CD T3 Cn C •H 15 0 iH Cn CL) > -H -P 3800 3400 3000 2600 2200 1800 1400 1000 150 100 cn T3 o •H CD ft CD CD M MH I •P tn o in 4H (0 50 3 c J L \ ' \ \ \ + + \ V3 \ - V . \> + \ \ \ \ \ \ . \ \ . 1000 2000 3000 E l e v a t i o n 4000 f e e t 5000 6000 Figure 3-3 Trends of f r o s t - f r e e period and cumulative growing degree days above 42°F for the Nelson area i t at higher elevations. In comparison, Brink's values f o r two stations near 2000 fee t show strong d i f f e r e n c e s : 97 days f o r Crescent V a l l e y and 144 f o r Nelson. Probably the broader v a l l e y and the warming provided by the lake at Nelson protect i t from cold a i r pooling and the f r o s t that occurs often at Crescent V a l l e y . Brink's estimate for Nelson i s about 3 weeks greater than indicated by the Canada Land Inventory Curve. Another important f a c t o r not shown i s the i n t e n s i t y of early f r o s t s , o since k i l l i n g may not occur before a temperature of 29 P i s reached, rather than at 32° F. Schmidt (1960) found f r o s t - f r e e period an unsatisfactory expression of l o c a l climate when estimating seedling growth. He regarded accumulated hour degrees above 43° F as u s e f u l to ind i c a t e the upper l i m i t s of the zone of Douglas-fir growth on the south face of Crest Mountain. Lacking t h i s i n -formation for source areas i n t h i s study, a s t r a i g h t l i n e was f i t t e d to values of growing degree days for stations at various a l t i t u d e s i n the Nelson area (Figure 3-3 ). The trend of t h i s measure, representing energy a v a i l -able for growth (above 42° F ) , decreased s i m i l a r l y to f r o s t - f r e e period. No heat threshold i s apparent, p o s s i b l y because the stations used are located i n d i f f e r e n t topographic s i t u a t i o n s . Since well-based estimates of growing degree days were not a v a i l a b l e for the Coastal areas, estimates of minimum f r o s t - f r e e period were read from Baker's (1944) curves and adjusted where l o c a l data showed them s t r i -k i n g l y d i f f e r e n t . The values appear i n Table 3- 4 . For both f r o s t - f r e e period and date of e a r l i e s t f r o s t , the Nelson 2000-foot s t a t i o n appears quite s i m i l a r to Coastal stations at comparable eleva^ t i o n s . S i m i l a r l y , the Nelson 3500-foot s t a t i o n i s very close to the values f o r the Seymour 3300-foot l o c a t i o n . For the Coastal c o l l e c t i o n s , a l l stations at s i m i l a r elevations show very close agreement i n both measures. 27 Table 3- 4 Estimated length of minimum f r o s t - f r e e period by area and e l e - vation of c o l l e c t i o n . Approximated from figures of Baker (1944) Area Elevation (feet) Period days Date of l a s t spring f r o s t Date of f i r s t f a l l f r o s t Nelson Haney Seymour Caycuse Paterson Lake 2000 3500 5200 100 500 1300 1800 50 400 1000 1900 2750 3300 600 800 133 105 58 187 178 156 138 189 180 165 134 119 109 175 170 11 18 May 15 June 8 Jul y 11 A p r i l 21 A p r i l 27 May May A p r i l 21 A p r i l 28 May May June June 15 A p r i l 29 May 3 8 20 7 October 6 September 19 August 27 November: 1 October 27 October 16 October 11 November 1 October 28 October 19 October 9 September 29 September 22 October 26 October 23 28 3.25 Growing-season p r e c i p i t a t i o n In addition to f r o s t - f r e e period and growing-season temperatures, plant populations i n each l o c a l i t y w i l l be adapted to the pattern of av a i l a b l e moisture during the growing season. This w i l l be modified by s o i l depth and porosity, topographic c h a r a c t e r i s t i c s of a breeding neighbourhood ("deme" of Langlet, 1971'a) and the amount of evaporation. The l a t t e r w i l l depend on the strength of i n s o l a t i o n and wind force: increases i n both increase water loss from plants and s o i l . S q u i l l a c e and S i l e n (1962) found growth of ponderosa pine (Pinus ponderosa Laws.) provenances i n plantations could be explained l a r g e l y by length of f r o s t -free period and percentage of annual p r e c i p i t a t i o n f a l l i n g during the growing season. No information on the net moisture a v a i l a b l e at the c o l l e c t i o n s i t e s was obtainable, so an attempt was made to estimate the proportion and the v a r i a b i l i t y of the t o t a l p r e c i p i t a t i o n received during the growing season. Growing season was that estimated for each l o c a l i t y i n Table 3-4. The period indicated was superimposed on the patternof annual p r e c i p i -t a t i o n derived from c l i m a t i c tables to obtain an estimate of the average growing-season p r e c i p i t a t i o n . This was expressed as a percentage of annual t o t a l . The values obtained appear i n Table 3-5. 29 Table 3-5 Mean growing-season p r e c i p i t a t i o n f o r areas of c o l l e c t i o n Growing-season precip'n C o l l e c t i o n area Weather s t a t i o n Sum % of Name Elev'n Name Elev'n (inches) annual Driest Precip'n month (inches) % of grow-season Nelson 2000 Nelson 1980 8.41 29.4 Jul y 1.23 14.6 Haney 500 Haney UBC 600 26.68 33.6 July 2.48 8.6 Seymour 100 Vancouver 15 14.50 23.5 July 1.04 7.2 Discovery Seymour 400 Seymour 583 33.61 23.6 Aug. 2.57 7.6 Creek Caycuse 600 Cowichan 580 17.40 21.0 Aug. 0.94 .5.4 Lake Paterson 800 Campbell 250 14.21 24.2 July 1.30 9.1 Lake River The Seymour Creek and Haney areas receive the greatest amount of p r e c i p i t a t i o n i n the growing season, and Nelson the l e a s t . The remaining areas receive nearly the same p r e c i p i t a t i o n . P r e c i p i t a t i o n received during the d r i e s t month varied strongly, from 0.94 inches f o r Cowichan Lake to 2.59 inches at Seymour Creek. The proportion of summer p r e c i p i t a t i o n received i n the d r i e s t month was highest f o r Nelson, although the amount of r a i n was f a i r l y low. Assuming these values represent one of the main s e l e c t i v e forces on the behaviour of the indigenous populations, the Haney and Seymour Creek material should require most water to grow s u c c e s s f u l l y and that from Nelson the le a s t . However, the driest-month proportion of growing-season p r e c i p i t a t i o n deviates more from the mean for Haney and Seymour Creek areas than f o r Nelson. 30 This probably w i l l exert pressure on those populations to accommodate t h i s v a r i a b i l i t y and to reduce t h e i r a c t i v i t i e s somewhat below those indicated by the mean values. The Caycuse material i n p a r t i c u l a r should be able to endure summer dry s p e l l s , since i t i s located only 10 miles west of the weather station and at the same elevation. These are rather crude d e f i n i t i o n s of growing-season..conditions,. since a unit of r a i n i n one area may accomplish less than the same amount i n a cooler area, or i n one with d i f f e r e n t topographic or edaphic conditions. An index of potential evapo-transpiration possibly would have more meaning to indicate the current forces shaping population be-haviour. Nevertheless, the data assembled indicate that differences i n temperature pattern and l e v e l , i n growing season length, and i n growing-season p r e c i p i t a t i o n amount and pattern exist between and at di f f e r e n t elevations within the seed source l o c a l i t i e s studied. I f these d i f f -erences exert selective pressures on the populations i n th e i r l o c a l i t i e s , the populations should exhibit differences s i m i l a r i n pattern to the major selective force when t h e i r offspring are reared i n common environment. 3.3 S t a t i s t i c a l analyses Wherever:: appropriate, the data collected was subjected to s t a t i s -t i c a l analyses involving analysis of variance (ANOVA) or covariance, or lin e a r regression. A l l are simi l a r i n t h e i r underlying assumptions: the individuals measured or judged are chosen without bias, the population, members are distributed normally about the mean, the variances of the popu-lati o n s being compared are not s i g n i f i c a n t l y different.: The l a t t e r re-31 quirement was tested using B a r t l e t t 1 s formula f o r t e s t i n g variance homogen-e i t y (Freese, 1967). Non-homogeneous data cannot be tested with f u l l confidence i n the r e s u l t s . I f appropriate, non-homogeneous data were transformed mathematically to improve t h e i r homogeneity. Confidence l e v e l s used generally were 5% and 1%, indicated by "*"-and "**", r e s p e c t i v e l y . 3.4 Cone morphology of mountain and western hemlocks 3.41 Introduction Structures used to separate taxa should possess features that are r e l a t i v e l y stable between environments and that allow quick and accurate determination of the pertinent feature by a number of independent workers. For that reason, f o l i a g e s i z e and shape are seldom s a t i s f a c t o r y characters to use i n i d e n t i f y i n g tree species. More commonly, plant taxa are i d e n t i f i e d l a r g e l y by the morphology of t h e i r flowers and f r u i t s . They are structures that remain f a i r l y consistent i n t h e i r composition and r e l a t i v e sizes over a broad range of habitats. However, studies on coniferous cones have revealed i n t r a s p e c i f i c v a r i a t i o n i n morphology with e c o l o g i c a l s i t u a t i o n of the parent tree (e.g. Tusko, 1963, S z i k l a i , 1969). Where strong d i f f e r e n c e s i n climate occur over a short distance, as i n mountainous areas, the e f f e c t s of environmental (e.g. temperature) gradients that e x i s t over a broad l a t i t u d i n a l b e l t can be found i n a short distance through di f f e r e n c e s i n a l t i t u d e . A number of analyses of cone morphology fromdifferent a l t i t u d e s i n a l o c a l i t y have been 32 conducted. Generally, cone dimensions decrease as a l t i t u d e increases (Velkov et <JLL. , 1965, Tusko, 1963, Barnes 1967, B i s s a l d i et a l . , 1970). An opposite trend of cone s i z e change with a l t i t u d e was found i n European l a r c h by Sindelar (1966). Yao (1971) found di f f e r e n c e s i n the trends of cone a t t r i b u t e s from Douglas-fir c o l l e c t e d over a great breadth of l a t i t u d e and a l t i t u d e . Depending on seed zone and feature, a l t i t u d e was p o t e n t i a l l y important to the v a r i a t i o n i n any feature. In a l l cases, tree-to-tree v a r i a t i o n i n the features studied was important. Intra-tree v a r i a t i o n i n cone features have been found i n a number of cases reviewed by S q u i l l a c e (1957) . S z i k l a i (1964) and Winjum and Johnson (1964) found that cone length of Douglas-fir d i f f e r e d between aspect i n the crown: southside cones were longer than those developed on the north side. Longer cones were found near branch t i p s (Winjum and Johnson, 1964). S z i k l a i ' s (1964) cone width responded s i m i l a r l y to cone length, except that i t showed s i g n i f i c a n t d i f f e r e n c e with l e v e l i n the crown. Nodal branches produced longer cones than internodal branches (Winjum and Johnson, 1964). S q u i l l a c e (1957) showed pos s i b l e metaxenial influence on cone c h a r a c t e r i s t i c s . Such e f f e c t s were found f o r cone length, scale s i z e : cone length r a t i o and scale width : scale length r a t i o . The only reported study of hemlock cone morphology i s by Taylor (1972). He found that cones of western hemlock varied i n length between 1.5 to 2.5 cm., while those of mountain hemlock were mostly between 4 and 7 cm., but included also some "intermediate" ones. 33 The evidence from these few studies shows that cone s i z e and scale proportions can vary within a tree because of exposure and n u t r i t i o n , but also because of p o l l e n parent i f one should be predominant. Con-sequently, study of such environmental and metaxenial-caused v a r i a b i l i t y should be conducted before features of cones of a species or region can be used to separate taxa or i n t r a - s p e c i f i c v a r i a n t s . In an attempt to characterize the nature and v a r i a b i l i t y of cones of western and mountain hemlocks, a small study was conducted. 3.42 Materials and methods Cones c o l l e c t e d from a si n g l e tree (E) at Totem Park were analysed for i n t r a - t r e e v a r i a b i l i t y . Cones from a l l trees included i n the i n t r a -s p e c i f i c crosses at Totem Park (Chapter 4) were analysed f o r the s t a b i l i t y of cone features i n an a l t e r e d environment ( i s o l a t i o n bags) and for meta-xenia. F i n a l l y , the most "stable" characters were used to t e s t f o r differences between mountain and western hemlock, and to determine aspects of i n t r a - s p e c i f i c v a r i a t i o n i n western hemlock. Sample s i z e - Number of cones The following procedure was used to determine sample s i z e : measure-ments of cone length and width were taken of 50 cones selected randomly from a well-mixed sample from one t r e e . Using a computer program, a random sample of these values was drawn one at a time, a f t e r which mean and standard deviation were c a l c u l a t e d . These c a l c u l a t i o n s were repeated a f t e r each new value was included i n the data u n t i l a l l 50 cones had been sampled. Pl o t s of the trend of standard deviation with sample s i z e were made and the approximate"'point: at -.which the standard d e v i a t i o n became r e l a t i v e l y stable and remained within about 5% of the f i n a l value was marked. The sample s i z e corresponding to t h i s point was determined. This process was repeated twice more for each measure, and the f i n a l sample s i z e used was determined from the average of the three. The r e s u l t s of a t y p i c a l run of the data appear i n Figure 3-4. I t i n d i c a t e s a lower sample s i z e of 40 cones, which was used i n subse-quent analyses. Sample s i z e - Number of scales per cone To determine the p o s i t i o n i n the cone from which scales should be taken, and the number of scales required per cone, i n d i v i d u a l scales of a number of cones were removed i n s e r i e s from base to t i p of cone and mounted f o r measurement. They were mounted by pressing them onto s t r i p s of masking tape stapled to s t i f f paper cards (Figure 3-5). The abaxial side of each scale remained i n view so that measurements of the bract and scale could be taken. The following measurements were taken fo r each scale, with the symbol indicated i n parentheses i n each case: length to mean p o s i t i o n of greatest width (L^), distance between point and t i p of bract (L,>), bract prong length (Br. P r . ) , bract length (Br.L.), bract width (Br.W.) and scale width (W). Scale length (L) was determined by t o t a l l i n g L^, and Br.L. The basic measurements are shown i n Figure 3-6. These were combined i n various ways to pro-duce a t o t a l of 26 measures per cone and scale. Measurements were obtained to 0.01 mm. by measuring the scales on an Addo-X tree-ring-measuring machine. The values recorded there on tape were transferred to computer cards f o r subsequent analyses. Several cones were dissected from base to t i p and each scale 35 E u 4-> ti Ci) B o u c o • r t +J (tJ •H > ft a -J W 0.6 0.5 0.4 0.3 K 0.2 0.1 Figure 3-4. \ N- A N/ N + + Cone length ++ + i + + ++ + ++ + + + .Cone width 0) I ft I <D S I N nJ 1 -H 10 20 30 40 S a m p l e s i z e c o n e s 50 Trend of standard deviation of cone length and width of western hemlock, and derivation of minimum sample size. 0.3 6 u • r t * 0) a 0 o 14-1 0.2 ° c o •rt -P • r t 0.1 | — o C T R E E 105 Figure 3-5. Scales of tree 105, mountain hemlock, mounted for measuring. w * 36 Br .Pr Br .W. Figure 3-6. I l l u s t r a t i o n of basic cone scale measurements. S c a l e n u m b e r f r o m c o n e b a s e Number 3-7. Trend of scale W-j- L from base of t i p of cone, western hemlock measured as ou t l i n e d . The r a t i o W A'. L was determined f o r each scale and p r i n t e d by computer. This r a t i o was chosen because inspection of scales of western and mountain hemlocks indicated that they d i f f e r e d s u b s t a n t i a l l y i n t h i s proportion i f the scales were taken from the middle of the cone. The r a t i o s determined f o r each scale i n a cone were inspected to e s t a b l i s h the trend within the cone. I t was found that the W T L r a t i o became f a i r l y stable near the upper middle of a cone of both species. This " s t a b i l i t y " extended over several scales of mountain hemlock cones, but only a few scales i n western hemlock (Figure 3-7 ). Since the sample s i z e indicated f o r s a t i s f a c t o r y standard error i n cone length was 40, sample s i z e of scales per cone was set a t one, making the sample per tree 40 a l s o . 3:43" - Results and:'discussion I n t r a - s p e c i f i c comparisons Intra-tree v a r i a t i o n The v a r i a t i o n of cone morphology within a tree was assessed on material from tree E, Totem Park, U.B.C. Cones were c o l l e c t e d from the t i p , middle and base of branches from the top, middle and lower parts of the open crown. The measurements were compared by s t a t i s t i c a l a n a l y s i s to detect any dif f e r e n c e s between c o l l e c t i o n l e v e l . Mean values per feature and c o l l e c t i o n stratum appear i n Appendix 1. Linear regression values f o r cone length (Cone.L), scale width (W), scale length (L), L^ and L^, and W f L from d i f f e r e n t crown l e v e l s appear i n Table 3-6. In general, dimensions are reduced f o r cones developed nearest the trunk. 38 Table 3- 6 Intra-tree v a r i a b i l i t y i n cone values. Linear regression values, branch t i p to base. Tree E , Totem Park. Sample size 40 Top Middle Lower Feature Slope 2 r S i g n i f . Slope 2 r S i g n i f . Slope r Signi: Cone L .127 .184 * * -.191 .377 * * -.133 .210 * * L -.202 .026 N.S. -.517 .154 * * -.327 .066 * * W -.241 .131 * * .387 .265 * * -.289 .193 * * L l -.473 .007 N.S. -.174 .101 * * -.191 .138 * * L2 -.022 .002 N.S. -.177 .061 * * .047 .005 N.S. W r L -.101 .026 N.S. -.673 .011 N.S. -.777 .014 N.S. The strongest trend of change always occurred i n the mid-crown area. Scale width and length show strong reduction, and W,4 L the l e a s t . A l -though the trends found follow a gradient from branch t i p to base, the 2 low c o e f f i c i e n t s of determination (r s) i n d i c a t e that much of the v a r i a t i o n i n these measures i s random. The cons-length r e s u l t s from t h i s tree agree with those of Taylor (1972) for western hemlock, and of Winjum and Johnson (1964) who found that longest Douglas-fir cones occur at the branch t i p s . Environmental influence As a check on the possible influence of the environment on cone and scale morphology, cones that had been enclosed i n p o l l e n i s o l a t i o n bags u n t i l harvested (Chapter 4) were measured as o u t l i n e d . These values were compared to those from unbagged cones from the same portions of the same trees. Features were tested s t a t i s t i c a l l y f o r d i f f e r e n c e s . 39 The mean value obtained by feature and parent tree and the r e s u l t s of the s t a t i s t i c a l comparisons of open versus bagged cones appear i n Appendix 2 . Cone length (Cone L.) was always increased by bagging, as were other measures of length: L 2 , bract length (Br.L.), bract prong (Br.Pr.) and scale length (L). The only exception was , the length from the scale t i p to the mean widest part. In no case was t h i s measure i n -creased s i g n i f i c a n t l y . Scale width (W) was a l t e r e d by bagging only on tree A, and then i t was reduced s l i g h t l y . However, bract width (Br.W.) was increased strongly i n bagged cones from trees A and E. The r a t i o s tested reacted as the measure placed i n the denominator: any length measurements there generally decreased the r a t i o f o r bagged cones. The response of the cones from tree D was l e s s than found f o r the others. No change i n means of L 2 and Br.W. were found, and the increases i n Br.Pr. and Br.L. were l e s s dramatic than those/from t r e e s A and E (pr o b a b i l i t y 0.05 vs. 0.01). The two simple measures that were not changed by bagging, W and L^, were selected f o r further a n a l y s i s because of t h e i r apparent s t a b i l i t y . These measures and the r a t i o W f were tested s t a ^ t i s t i c a l l y to compare western hemlock populations from s i m i l a r elevations i n the Coastal and I n t e r i o r portions of the range. The c o l l e c t i o n s compared were those from Haney at 1800 feet, Seymour at 2750 feet and Nelson at 2000 fee t above sea l e v e l . The r e s u l t s are presented i n Table 3-7 . 40 Table 3-7 Comparison of western hemlock cone scale measures from Coastal and Interi o r origins - mean values i n mm. C o a s t a l Seymour Haney Parent Parent I n t e r i o r Nelson Parent Tree W W -r L '. Tree W W -r L Tree W W T L± 96 7.15 2.16 85 6.80 2.36 11 6.46 1.70 97 6.69 1.86 86 6.63 2.20 12 6.86 1.93 98 6.46 2.25 87 6.15 2.16 13 7.18 1.99 99 6.97 1.79 88 6.53 2.39 14 6.96 1.78 100 7.25 2.42 15 7.39 1.97 101 6.51 2.09 102 7.62 2.40 Mean 6.95 2.14 6.54 2.28 6.97 1.85 Scale width of the Haney material was d i f f e r e n t from that of the other c o l l e c t i o n s , which did not d i f f e r . The difference was not con-sis t e n t , however, as the mean values for individuals from d i f f e r e n t populations overlapped. For instance, trees 101 from Seymour and 88 from Haney had almost i d e n t i c a l means, and trees 98 from Seymour and 11 from Nelson gave means of 6.46 mm., far below the o v e r a l l mean for the c o l l e c t i o n . The s i m i l a r i t y between the Seymour and Nelson means shows that scale width alone does not indicate any difference between the cone morphology of Coastal and Int e r i o r populations. The fr a c t i o n W f gave dif f e r e n t r e s u l t s . Both Coastal col l e c t i o n s d i f f e r e d from the Nelson value. But the Seymour and Haney values d i f f e r e d from each other also, and i n a l l cases, parent treedri." i area was a s i g n i f i c a n t source of variance, eliminating t h i s feature as a s e n s i t i v e means of separating Coastal and I n t e r i o r cones from si m i -l a r elevations. I n t e r - s p e c i f i c comparisons In order to f a c i l i t a t e the determination of putative hybrids of western and mountain hemlocks, and i n view of the d i f f e r e n c e s i n cone scale features apparent i n Plates 1 and 2, measurements of cones and scales were taken from samples of each species and compared. The western hemlock material measured or i g i n a t e d from both Coastal and Inter areas (Nelson, Haney and Mount Seymour) at about 1900 feet above sea l e v e l ( a . s . l . ) . Mountain hemlock cones from 3000 fee t a . s . l . on Mount Seymour and from 3500 feet near the Hope S l i d e provided the data for that species. The mean, standard deviation and s i g n i f i c a n t d i f f e r -ences between the species appear i n Appendix 3\ S i g n i f i c a n t d i f f e r e n c e s between the two species were found i n every feature. However, the following appear to be the most us e f u l measures because of the strong d i f f e r e n c e between the mean values and t h e i r low standard deviations: Cone length, z,^, W T L, r L, L •: Br.L., and number of scales per cone, and 7 L^. and scales per cone i n p a r t i c u l a r , showed very strong d i f f e r e n c e s between the two species. Since scale number i s l i k e l y to vary with cone length, regressions of scale number with cone length were ca l c u l a t e d f o r the two species and compared (Table 3-8 ). 42 Table 3- 8 Comparison of regressions of the number of cone scales with cone length for western and mountain hemlocks . 1 2 Species n Slope, Intercept, Stan. r S i g n i f . scales scales error, per cm. scales  W.H. 240 5.59 15.73 2.63 0.404 * * M.H. 151 8.27 49.67 6.24 0.524 * * "'""W.H." western hemlock, "M.H.", mountain hemlock. Covariance comparison of the regressions showed that they d i f f e r e d i n slope and intercept. In both cases, the value f o r mountain hemlock was highest. These t e s t s showed that the species d i f f e r i n the structure of t h e i r cones, so that cones of equal length w i l l s t i l l d i f f e r strongly i n number of scales. A feature not expressed i n the measurements i s the nature of the bract prongs of the species. That of mountain hemlock i s mucronate (Plate 2 (11), while that of western hemlock i s much blunter, often developing no further than a small bump (Plate 1 (10 and 11). This i s r e f l e c t e d i n the c o e f f i c i e n t s of v a r i a t i o n (S.D. f mean) f o r the species: 15.9 percent f o r mountain hemlock and 40.7 percent f o r western hemlock. The v a r i a b i l i t y of thismeasure i n western hemlock made any of the r a t i o s using Br.Pr. as denominator so v a r i a b l e that the variances ("Bartlett -.test s i g n i f i c a n c e " values) of the two species were not d i f f e r e n t (Appendix 3 ) whereas they were d i f f e r e n t f o r nearly a l l other measures. Figures 3-8 and 3-9 present p l o t s of the r e l a t i o n s h i p between W and L and L 9 and L. The d i f f e r e n c e s between the two species are 43 11 U 10 9 8 -P • r t o western hemlock A mountain hemlock _L A A A *A 8 9 10 11 S c a l e l e n g t h mm. 12 13 Figure 3-8. Scale W : L r e l a t i o n s h i p for i n d i v i d u a l mountain hemlock and western hemlock cone parents. 6 5 4 3 2 1 0 -1 western hemlock A mountain hemlock Figure 3-9. 8 9 10 11 S c a l e . . l e n g t h mm. 12 13 Scale L 2 : L r e l a t i o n s h i p for i n d i v i d u a l mountain hemlock and western hemlock cone parents. 44 apparent. Although scale lengths overlap broadly, L 2 f o r mountain hemlock i s lower than f o r western hemlock. Trees 109 and 167 had negative values f o r L 2 because the bract t i p extended above the widest part of the scale. Scale width i s much greater i n mountain hemlock, producing a W -7 L r a t i o of 1.00 (Appendix 3 ), rather than 0.62 for western hemlock. Comparing the f i g u r e s , i t i s c l e a r that L 2 p l o t t e d over L gives c l e a r e r separation of the species than W over L. In both cases, the pure mountain hemlock trees from Mount Seymour, numbers 105-109, were strongly separated from a l l western hemlock trees. Yet W over L placed tree 102, a western hemlock from 2750 feet on Mount Seymour, near tree 183, a mountain hemlock from 3500 fee t near Hope. The same trees were separated c l e a r l y when L 2 was p l o t t e d over L. Tree 181 from the Hope c o l l e c t i o n behaved i n c o n s i s t e n t l y : W over L placed i t among the mountain hemlocks whereas L 2 over L placed i t nearer western hemlock. I t and i t s c l o s e s t neighbour i n the L 2~over-L p l o t , tree 99, were compared (Table 3 - 9 ) . They d i f f e r e d strongly i n mean cone length, since each was some-what shorter than general f o r i t s species. Scales of tree 99 were of average s i z e , but they had a much lower point of maximum width and a longer bract prong than customary f o r western hemlock, g i v i n g i t a higher value of and reducing that of L 2 . Tree 181 appeared to be a t y p i c a l mountain hemlock with the exception of s l i g h t l y shorter bracts and higher point of maximum scale width, producing a larger L,>. Features of i t s bracts: mucronate t i p and broad, high "shoulders", were t y p i c a l 45 ofmountain hemlock. Values f o r other features of bracts and cone scales from these trees were checked. Both trees gave values generally d i f f e r e n t from the mean values f o r the species shown i n Appendix 3 and close to each other (Table 3-9 ) . Table 3- 9 Comparison of selected cone and scale values f o r trees 99 and 181 to mean values f o r western and mountain hemlocks. 2 (values, apart from Cone L in mm or mm ) Feature No. Symbol Western hemlock Mean Tree 99 Tree 181 Mean S.D. Mean S.D. Mountain hemlock Mean 1 2 3 4 5 6 8 9 Cone L(cm) W L l L2 Br.L. Br.Pr. L W t L 11 W 12 L 2 f L 15 W * Br.W. 2.08 6.97 3.54 4.53 3.19 0.78 11.27 0.62 2.00 0.40 2.57 17 Br.W *• Br.Pr. 10.04 19 L -T Br.L. 3.64 22 W (L) 78.99 23 Br.L. (Br.W.) 8.89 1.74 6.97 3.92 2.93 3.87 1.01 10.77 0.65 1.79 0.27 2.51 2.93 2.80 75.30 10.97 0.19 4.24 0.44 9.76 0.39 3.17 0.60 2.98 0.42 4.89 0.26 0.75 0.82 11.05 0.04 0.88 0.15 3.13 0.05 0.27 0.13 2.63 0.77 5.27 0.21 2.27 8.99 108.11 1.59 18.26 0.56 0.71 0.47 0.74 0.49 0.18 0.75 0.06 0.41 0.06 0.22 1.27 0.21 12.89 2.32 5.03 9.83 3.67 0.59 5.55 1.08 9.84 1.00 2.71 0.06 3.25 2.86 1.78 97.13 17.01 46 A b r i e f check of the number of scales per cone was made f o r these trees. They d i f f e r e d strongly and f e l l within the l i m i t s established already f o r the separate species. In view of Taylor's (1972) observation of the cone lengths of the two species, p l o t s of mean number of scales with cone length were made f o r several trees of each species (Figure 3-10 ). I t i l l u s t r a t e s the strong d i f f e r e n c e s i n these features between western and mountain hemlocks. Although there i s v a r i a t i o n , the features are f a i r l y stable and separate the species c l e a r l y . The l i m i t s i n i n d i v i d u a l cone length found are: western hemlock, 0.8 to 2.8 cm.; mountain hemlock, 1.7 to 6.9 cm. They agree generally with Taylor's l i m i t s of 1.5 to 2.5 cm. f o r the former and mainly 4.0 to 7.0 cm. f o r the l a t t e r . Values between 2.5 cm. and 4.0 cm. he considered i n d i c a t i v e of mountain hemlock. Mean lengths varied from 1.17 to 2.54 cm. i n western hemlock and 3.77 to 6.08 cm. i n mountain hemlock. In each case, the smallest means came from trees located at about 3500 feet i n the Hope S l i d e c o l l e c t i o n . The trees g i v i n g the longest cones of mountain hemlock were located at 3300 feet i n Mount Seymour Park, and those of western hemlock at 2000 feet near Nelson. C o e f f i c i e n t s of v a r i a t i o n of mean cone length by species are 14.3% and 16% for western and mountain hemlocks, r e s p e c t i v e l y . Thus, they are about equally v a r i a b l e . Despite the number of d e t a i l e d measurements taken on cone scales and bracts, i t appears the best features to use i n separating western and mountain hemlock cones are cone length and number of scales combined, ei t h e r as a value of scales per cm. of length or i n a f i g u r e r e l a t i n g the two (Figure 3- 10 ) . 47 100 gj 80 o CJ u cu C4 cn cu i-i rd CJ cn m O 60 & 40 20 108 107 A A 109 168 A 167 172 of sepa ; 106 A 105 169 171 174' g h 170* '173 I l_ 166 _L 4 5 C o n e 6. 7 8 l e n g t h cm. 10 11 Figure 3-10. Trend of scale number with cone length for western and mountain hemlocks. Figure 3 -11. Comparison of " t y p i c a l " scales of western and mountain hemlock cones. Values i n mm. 48 One tree (172) d i f f e r e d strongly from the other species when cone length and scale number were re l a t e d (Figure 3-10) . Number of scales was almost h a l f way between those of the t y p i c a l species, while cone length tended toward that of western hemlock. The r a t i o of L 2 to L (-0.001) placed i t well within the l i m i t s f o r mountain hemlock. S i m i l a r l y f o r W -5- L (0.908), Br.W. * L (0.626), L -5- Br.L. (1.627), L^ 4- L 2 (40.33) and L^ + bract area (0.228). Bract shape and t i p also are those of mountain hemlock (Figure 3-11) . Despite the reduced number of scales, the cones from tree 172 appear to be t y p i c a l of mountain hemlock i n a l l other respects. I f i t was from a hybrid between western and mountain hemlocks, as the number of scales per cone suggests, the other cone features considered diagnostic above, p a r t i c u l a r l y W 4- L, L^ 4- h^, L 2 4- L, L 4- Br.L. and nature of bract prong, should have shown some intermediacy a l s o , assuming that each feature i s c o n t r o l l e d by a number of genes and those from mountain hemlock are not dominant. Roche's (1967) r e s u l t s , summarized above, i n d i c a t e that a number of genes are involved i n c o n t r o l l i n g cone scale shape i n white and Engelmann spruce. Furthermore, the v a r i a b i l i t y i n cone a t t r i b u t e s he found i n d i c a t e s that the genes are somewhat independent, producing an i r r e g u l a r range of types from the areas intermediate i n elevation to the more t y p i c a l parental types. Probably intermediacy would occur i n the cone c h a r a c t e r i s t i c s of a hybrid between the native hemlocks, p a r t i c u l a r l y the f i r s t - g e n e r a t i o n hybrid. I f i t was successful and reached reproductive age, rearrangement of the genes during meiosis would increase the breadth of the genetic combinations c o n t r o l l i n g cone c h a r a c t e r i s t i c s , tending to produce v a r i a b i l i t y i n i t s o f f s p r i n g of the sort found by Roche. Taylor (1972) found evidence of introgression between western and mountain hemlocks i n the chemical intermediates detected. No features such as those studied here were reported on, so i t i s not possible to test the s t a b i l i t y of the cone features considered diagnostic when a hybrid i s produced. But considering the intermediacy of cone morphology i n many pine species, egs. Pinus attenuata x radiata, P_. ponderosa x j e f f r e y i , P. banksiana x contorta ( L i t t l e and C r i t c h f i e l d , 1969), cone morphology i n pines apparently i s governed by many genes, usually without dominance. Assuming the same would occur i n a hybrid between the native hemlocks, i t i s concluded that tree 172 i s not a hybrid between the species; despite i t s reduced number of scales per cone, the other cone attributes confirm i t as a mountain hemlock. 3:4.4 Summary of cone studies These studies have established a few facts about the morphology of western and mountain hemlock cones. Mountain hemlock consistently had more scales per cone and per cone centimeter, broader scales, much longer bracts r e l a t i v e to scale length, a larger and more mucronate bract, a higher W r L and L^ f L 2 r a t i o , a lower L 2 f L r a t i o , and longer and more consistent bract t i p s r e l a t i v e to bract length and bract width than western hemlock. Cone length also was greater and scale length shorter than i n western hemlock, although these features are less r e l i a b l e than the foregoing. No consistent difference i n morphology was found for c o l l e c t i o n s of western hemlock from Coastal vs. I n t e r i o r areas at the same elevation when the most stable features were used. Strong intra-population v a r i a b i l i t y i n scale width and W 7 L were noted for the Coastal and Inter i o r c o l l e c t i o n s examined. The " t y p i c a l " western hemlock cone scale i n t h i s material (Figure 3-11) resembled one of Sargent's (19 00) (Plate 1 (11) and not the other (Plate 1(10). Only a few scales shaped l i k e the l a t t e r were found. However, the cones studied here (Figure 3-11) generally had smaller scales than those depicted by Sargent (1947), a l -though cone lengths were near the middle of the ranges considered t y p i c a l (Table 2-\ ). Some of t h i s v a r i a t i o n might be due to differences i n environment about each tree, but such strong differences i n cone size were noted between trees growing next to each other that much of the difference l i k e l y was genetic. The chief difference between t h i s material (Figure 3-11) and Sargent's (Plate 2) l i e s i n the shapes of the cone scale and bract. Sargent's i s more elongated than these, and the bract i s only s l i g h t l y shouldered. Forty cones was established here as the minimum to maintain a stable estimate of cone length. Sample size could be reduced somewhat i f cones were collected from one portion of a branch, rather than from the entire upper crown, since cone and scale dimensions generally decreased from branch t i p to base. Environment about the cones can affect the dimensions markedly, and r a t i o s less so, although s i g n i f i c a n t l y . Cones collected to analyse the v a r i a t i o n pattern between populations of western hemlock should be collected from the exposed portion of tree crowns and from trees growing as nearly as possible i n si m i l a r environments, or across the breadth of environments i n each l o c a l i t y to reduce the chance that results are strongly affected by a small area of a t y p i c a l environment. Judging by the differences between bagged and "open" cones, cones collected i n d i f f e r e n t years could r e f l e c t strongly the differences i n weather pattern, p a r t i c u l a r l y temperature, rather than only differences i n cone char a c t e r i s t i c s between populations sampled. Such strong differences i n cone and scale morphology were found between mountain and western hemlocks that t h e i r maintenance as separate species seems warranted. Stebbins (1972) commented that "...the great majority of the differences between ecotypes are controlled by many genes." I f western and mountain hemlocks were merely ecotypic separates from one parental precursor, a clear trend of forms uniting them l i k e l y would be evident where they occur sympatrically. No such trend was found i n the l i m i t e d samples studied here. Taylor (1972), looking for intermediates, found far fewer chemical "hybrids" than indicated by morphological rating based both on foliage and cone measurements. Apparently, hybridization can occur between the species, but r a r e l y . Where any putative hybrids are located i n future, t h e i r cones can be analysed using the r e l a t i v e values of scale number per cone found here, the r a t i o of scale length to width, the r a t i o s of to and Ii^ to L. In addition, the nature of the bract prong (mucronate or blunt) should be included. Assuming a number of genes, each with equal action, control each feature i n both species, and no dominance or epistasis occurs, the F^ hybrid should produce cones giving values intermediate to those found for the species (Appendix 3 ), plus bract prongs more regularly formed than in western hemlock, and less mucronate than for mountain hemlock. 53 3.5 Nursery studies of mountain and western hemlock populations Two nursery t r i a l s were established i n an e f f o r t to detect d i f -ferences i n seedling growth and performance due to species, c o l l e c t i o n area and i t s climate, and parent tree. 3.51 Container nursery study Preparations began i n March bf 1970 for the production of plantable seedlings from the 1968 population c o l l e c t i o n to permit measurements of "full-grown" seedlings and the establishment of a f i e l d t r i a l from which early estimates of provenance and within-provenance va r i a t i o n i n seedling dimensions and performance could be obtained. The seedlings were raised at the P a c i f i c Forest Research Centre of the Canadian Forestry Service i n V i c t o r i a , B.C. Materials and methods Seed from each tree had been extracted, de-winged and cleaned by hand. Empty seeds were blown off on a forced-air column separator (Silen, 1964a) following determination of the suitable current for each seed l o t by checking the quality of the separation on the fluorescent screen of a "Softex" X-ray machine. Empty seeds were discarded without being counted and the " f i l l e d " seeds were returned to cold storage u n t i l needed. Samples containing 100, 150 or 200 wole seeds were counted out for 106 parent trees - a l l those for which s u f f i c i e n t seed was available. These were soaked i n tap water at room temperature, the seeds being held submerged i n the water by a wad of cotton wool. After 48 hours the water was sucked out, the v i a l s were stopped up loosely by the cotton wool, and the seed was placed i n the cold room ( 1° + 1° C) for 21 days of s t r a t i f i c a t i o n . Batches of seed were soaked and put i n to s t r a t i f y on successive days to spread out the work of sowing while keeping the s t r a t i f i c a t i o n period constant. Seeds were sown onto a 3:1 (volume:volume) peat-vermiculite mixture i n styrofoam moulds designed by the Canadian Forestry Service for the production of "plug-grown" seedlings. Two or more seeds were sown i n each plug. Then they were covered with chipped granite g r i t to act as a mulch. Water was sprinkled over every tray as soon as the seed was sown. A plug block consisted of four compartments, each of which con-tained 48 moulds f i l l e d with rooting medium. The cone parents were assigned randomly to the compartments i n a block, permitting a maximum of 48 seedlings per family following thinning to one seedling per cavity Parent tree number was marked on each compartment. The sowings were not replicated. Sowing was conducted between A p r i l 5 and 7, 1970. Loaded blocks were stacked under a p l a s t i c sheet i n the greenhouse at U.B.C. to keep them as moist as possible before trans-shipment to V i c t o r i a . On A p r i l 9 1970, a l l blocks were transported to the P a c i f i c Forest Research Centre of the Canadian Forestry Service i n V i c t o r i a , where the seedlings were tended. Each block was soaked again before being placed i n a mist. o o room to incubate. Conditions i n the room were: temperature 22 +_ 2 C, humidity approximately 95%, l i g h t 16 hours per day. When the germinants were v i s i b l e above the g r i t , the blocks were removed from the mist room and placed outside under natural conditions 55 except for diffuse overhead shade. Nylon screening well above the seed surface transmitted 46 per cent of f u l l l i g h t . On May 6, 1970, ten days after the blocks were moved outdoors, some damage to seedlings was noted i n certain blocks. The germinants were pale and shrunken, with collapsed cotyledon t i p s . Less severely damaged germinants had cotyledens with pale areas at the base or along the margins. Damage was confined mainly to single blocks containing seed from four parent trees, rather than occurring only by seed parent. The f l a t s were checked for pathogens by a pathologist on the s t a f f of the Centre, but none was found to which t h i s kind of damage could be attributed. Two warm, dry days had occurred shortly after the blocks were placed outside and i t seems probable that the succulent germinants had suffered drought damage after being moved out-doors d i r e c t l y from the mist room. The blocks were moved indoors to the greenhouse, where they remained the rest of the summer. Preparations were begun immediately to re-sow the seed l o t s showing losses. These replacement sowings were done into paper cups and placed into the mist room after soaking and s t r a t i f y i n g as before... After about three weeks i n the mist room the seedlings were transplanted where needed. Transplantings began July 2, 1970, and re-quired two days. In the meantime, the surviving seedlings from the f i r s t sowing were being raised under the following conditions: temperature 21-23° C, l i g h t -18 hours per day of mixed white flourescent and incandescent bulbs, humidity - variable. Watering was done two to three times weekly as required. No f e r t i l i z e r was applied before the germinants had exposed cotyledons. I t was injected into the watering system at a rate s u f f i c i e n t to produce a concentration of f i r s t 50 ppm, then 100 ppm, of nitrogen. Details of the c u l t u r a l regime used are presented i n Appendix 4 . Water was sprinkled onto the trees for half an hour after each f e r t i l i z i n g to wash the f e r t i l i z e r solution from the tops and prevent any "burning" caused by high mineral concentrations on the foliage and tender stems. A weekly water drench was used to flu s h the blocks out and reduce mineral s a l t buildup. Once transplants had been placed i n the blocks, the f e r t i l i z e r con-centration was halved to reduce the chance of simi l a r damage to the younger seedlings. I t was increased to 100 ppm of nitrogen once the youngest seedlings were large enough. On August 25, 1970, the f e r t i l i z e r used was changed to a 15-15-30 NPK formulation and watering was reduced somewhat to discourage further height growth and to hasten bud "set". Watering was reduced to once weekly i n October. A l l blocks were placed randomly i n one section of the greenhouse. Their positions were changed at irregular intervals to reduce the i n -fluence of differences within the greenhouse on seedling development. When the plants had apparently set buds the blocks were moved outdoors under l i g h t shade. They were transported i n a refrig e r a t o r truck to the U.B.C. Research Forest near Haney, where they remained outside without shade over winter. Shortly afterward, measurements were conducted on the progeny from each seed parent. The following values were obtained: height (epicotyl length i n cm. to the base of the terminal bud), stem diameter i n mm. at the cotyledons, the number of branches, the length i n cm. of the longest branch. These were taken from a random sample from each family, although 57 trees that had been transplanted into the f l a t s after re-sowing were excluded. In addition, the number of surviving trees and the number exhibiting damage attributable to f r o s t were t a l l i e d . The size of sample measured for height, diameter and branches was determined by the number of families available for each area. Where f i v e or fewer families survived, 15 trees per family were sampled. Where six or more families were a v a i l -able, 10 trees per family constituted a sample. A l l the data were analysed by computer. In May of 1971 seedlings from each family were planted i n the U.B.C. Research Forest at Haney i n a l i n e - p l o t design replicated twice to provide short-term observations on the behaviour of the progeny. Results Height The means and standard deviations of the data grouped by region and elevation appear i n Table 3-10. Mountain hemlock was much shorter than a l l western hemlock popu-la t i o n s , even that from 2750 feet on Mount Seymour. In each case, B a r t l e t t ' s t e s t showed variance was non-homogenous for western hemlock data from at least one of the areas. Since height and diameter were the most important expressions of growth taken, attempts were made by combining them and transforming them to remove the s i g n i f i c a n t non-homogeneity. None succeeded for height and diameter separately, but no s i g n i f i c a n t non-homogeneity existed for the height: diameter r a t i o when the Nelson, Haney and Seymour co l l e c t i o n s were compared. But the variances from t h i s r a t i o were different between elevations within the Nelson and Haney material, weakening tests on trends within these areas. Table 3-10 Dimension summary of plug-grown seedlings-from 1968 population collections. Longest Eleva- No-. Height Diameter Height/Dia, Branch Branch Area tion meas- (cm.) (mm.) (cm./mm.) (cm.) Number (ft.) ured Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean • S.D. Western hemlock Nelson 2000 60 14. .77 3, .51 2, .47 0. 48 6, .02 1, .10 6. 21 2. 01 6, .73 1, .70 3500 60 13. ,98 2, .43 2. .66 0. 36 5, .29 0, .82 6. 86 2. 27 6. ,45 1, .63 5200 42 10. .12 4, .33 1, .88 0. 56 5. ,20 1, .52 4. 88 2. 87 5. .33 1. .77 Haney 100 90 16, .48 3. ,89 2, .67 0. 49 6. ,22 1, .26 7. 07 2. 47 8. ,25 2. .51 500 120 14. .23 3. .64 2. .45 0. 48 5. .84 1, ,23 6. 09 1. 71 6. ,88 2. .03 1300 120 14. .55 3. .46 2, .68 0. 48 5. .47 1. .06 5. 71 1. 70 7. ,70 2. .42 1800 100 12. ,93 2. .78 2, .68 0. 53 4. .93 1. .05 5. 34 1. 58 6. .83 1. .87 Seymour 50 90 15. ,65 3. .90 2. .71 0. 50 5. ,80 1. ,05 6. 37 1. 78 7. ,19 2. .11 400 80 13. ,53 3. .45 2. .54 0. 47 5. .35 1. ,09 5. 76 2. 03 6. .36 1. ,80 1000 73 14. ,36 3. ,12 2. ,70 0. 40 5. ,37 1. ,08 7. 10 2. 09 7. ,22 1. ,77 1900 75 14. ,03 3. .90 2. .77 0. 52 5. ,10 1. ,06 5. 94 2. 05 7. ,51 2. ,20 2750 70 12. ,86 3. ,04 2, .48 0. 35 5. .20 1. ,12 5. 33 1. 66 6. .84 1. ,68 Caycuse 600 90 15. .16 3. .12 2. ,86 0. 54 5. ,39 1. ,05 6. 62 2. 08 8. ,59 2. .10 Paterson 800 75 15. ,10 3. .16 2. ,55 0. 44 5. ,97 1. ,06 6. 32 2. 09 8. .03 2. .01 Lake Mountain hemlock Mount Seymour 3300 22 7.61 1.83 2.76 0.38 2.77 0.60 3.00 1.05 6.27 1.83 59 Consequently, the v a l i d i t y of the analyses and linea r regressions presented here i s weakened by t h e i r f a i l u r e to s a t i s f y a basic re-quirement of such analyses. Prom ANOVA of the height measurements from a l l areas represented: Nelson, Haney, Seymour, Caycuse and Paterson Lake, both area and cone parent were important components of the v a r i a t i o n i n height (P less than 0.001 i n a l l cases). Further analysis of the data from only those l o c a l i t i e s including c o l l e c t i o n s from di f f e r e n t altitudes (Nelson, Haney and Seymour) revealed a strong influence of elevation within region. Generally, seedling mean height decreased as the a l t i t u d e of seed o r i g i n increased. This was tested separately for each of these three areas by linea r regression, f i t t i n g a l i n e of the form: Height = a + b (Elevation). The res u l t s appear i n . Table 3-11 . Table 3-11 Regressions of seedling height (cm';) on elevation of seed source (thousands of feet) No'e -of-' Limits*.of Sig-Seedlings Al t i t u d e Standard ^ n i f i -L o c a l i t y Measured (ft) • Intercept Slope Error r cance Nelson 162 2000- 18.06 -1.41 0.217 0.21 * * 5200 Haney 430 100- 15.92 -1.52 0.265 0.07 * * 1800 Seymour 388 50- 14.96 -0.70 0.184 0.04 * * 2750 "'"Coefficient of determination = proportion of t o t a l variance accounted for by the regression. 60 A highly s i g n i f i c a n t relationship existed i n each case, but the proportion of the va r i a t i o n removed by the regression was low, ranging 2 from 4 per cent (r = 0.04), to 21 per cent for the Nelson c o l l e c t i o n . This regression i s based on the fewest trees and represents the greatest range of alt i t u d e s (3200 feet) and so might be expected to r e f l e c t a strong response i n height growth i f i t i s under selective pressure at t h i s stage of seedling development at high a l t i t u d e s . The regressions were compared by covariance analysis. S i g n i f i c a n t differences i n slopes and intercepts were found between the regression l i n e for Seymour pro-venances and the other two, which did not d i f f e r . The weakness of the re l a t i o n i n the other l o c a l i t i e s and the high significance of the cone parent i n the analysis of variance indicated that i t was a strong factor affecting seedling height to t h i s stage. A comparison was made between 100-seed dry weight and mean seedling height for 23 parent trees (Figure 3-12). The r e l a t i o n was not s i g n i f i -2 cant (r = 0.006, d.f. 22). Seed factors other than weight, which r e f l e c t s mainly "endosperm" weight and embryo si z e , that might af f e c t seedling growth are embryo and "endosperm" maturity. Allen (1958a) noted strong differences i n seed maturity between parent trees of western hemlock when the seed was picked early. Since a l l trees sampled here were picked at nearly the same time i t i s probable that differences i n seed maturity existed between trees i n a stand, and that some of the differences between the development of the progeny, such as germinative rate and general vigour, are due i n part to such differences i n seed maturity (Gustafsson and Simak, 1956). X-ray plates were made of seed samples from a few trees 61 20 19 .18 17 16 15 14 13 12 i i u 10 .77 1 •79 62 69 49 70 80-18 32 53 98 . 81 •21 •7.6 _ I _ .25 58 17: 51 38 82* ; * 153 35 39 r = 0.006 N.S. .14;.. .15 .16, .17 M e a n 1 0 0 .18 .19 .20 .21 ,22. s e e d w e i g h t gm. .23 .24 Figure 3-12 Relationship between mean seed weight and mean seedling height for 23 parent trees. 62 collected early to check on seed development. Incompletely developed seed were apparent from two trees, yet generally i n low proportion. Mean heights of the seedlings are: tree 11, 16.57 cm; tree 14, 11.94 cm. Thus, although differences i n seed maturity could be p a r t l y responsible for seedling growth differences, i t appears that genotypic differences are most important. Mean heights for d i f f e r e n t elevations within the Nelson, Haney and Seymour areas were compared. For each l o c a l i t y , s i g n i f i c a n t differences were found between populations at the lowest and at the highest a l t i t u d e . The greatest d i f f e r e n t i a t i o n appeared i n the Haney c o l l e c t i o n , where s i g n i f i c a n t d i f f -erences were found between a l l but the col l e c t i o n s from 500 and 1300 feet. From Nelson, the highest c o l l e c t i o n (5200 feet) was d i f f e r e n t from the others (2000 feet and 3500 f e e t ) . The pattern from Mount Seymour was i r r e g u l a r . S i g n i f i c a n t differences occurred between the low population (50 feet) and the highest (2750 f e e t ) , but also between the 50-foot and the 400-foot co l l e c t i o n s . The mean seedling heights for these l a s t c o l l e c t i o n s are, i n order, 15.65 cm. and 13.53 cm. The means are based on 90 and 80 seedlings, representing 9 and 8 parent trees, respectively. Since the mean seedling height from the 400-foot area was less than those from the 1000 and 1900-foot c o l l e c t i o n s , the means of the indi v i d u a l trees from the 400-foot elevation were examined. None of the 8 parents gave an inordinately low value that depressed the population mean, so the l a t t e r figure seems a true r e f l e c t i o n of the c o l l e c t i o n , and l i k e l y represents the population mean adequately. One-hundred seed weights of s i x of these trees were available for comparison with those of only two parents (trees 69 and 70) from the lower c o l l e c t i o n . Seed weight varied between 0.148 and 0.215 gm. per 100 seed for the 400-foot c o l l e c t i o n , whereas those from the lower c o l l e c t i o n were near 0.18 gm. per 100 seed. One family, (tree 69) was s i g n i f i c a n t l y t a l l e r than a l l the re s t , indicating again that seed weight was not an important factor influencing height under these conditions. The mean heights of the Vancouver Island collections were compared to those from similar elevations on the Lower Mainland. The eaycuse mean of 15.16 cm. was not s i g n i f i c a n t l y d i f f e r e n t from that of Haney at 500 feet (14.23 cm.), but i t was greater than the mean for the 400-foot c o l l e c t i o n from Mount Seymour just discussed. The Paterson Lake c o l l e c t i o n (800 feet elevation) gave a mean height of 15.10 cm., which was not s i g n i f i c a n t l y different from the mean obtained from the Seymour 1000-foot c o l l e c t i o n (14.36 cm.). Thus, i n height alone, these Vancouver Island provenances do not appear to d i f f e r from those from the Lower Mainland, with the exception of the anomalous one from low on Mount Seymour. Arranging a l l cone parents by mean height of progeny, i t was clear that great overlap existed between c o l l e c t i o n l o c a l i t i e s (Appendix '5 ). For example, the c o l l e c t i o n from the base of Mount Seymour produced o f f -spring ranging from near the mean (tree 71 at 13.7 cm.) to near the t a l l e s t (tree 65 at 19.0 cm.). Progeny originating from the top of Mount Seymour were grouped i n the lower half of the array, yet three trees (97, 98 and 100) produced offspring no di f f e r e n t from the ove r a l l mean. Similar to the trend from the low-elevation Mount Seymour c o l l e c -t i o n , -the trees from the lower Haney c o l l e c t i o n , at approximately 100 feet elevation, ranged from near the mean (trees 18 and 21) to the t a l l e s t seedlings (trees 20 and 25). Despite i t s smaller sample of parent trees, the c o l l e c t i o n from the lowest elevation (2000 feet) i n Nelson produced a broad range i n heights, between 11.94 cm. and 16.57 cm. This compares to 64 the d i s t r i b u t i o n of the means of the seedlings from a similar a l t i t u d e (1900 feet) on Mount Seymour, but i s d i f f e r e n t from that for the Haney c o l l e c t i o n o r i g i n a t i n g at 1800 feet. I t i s clustered i n the lower t h i r d of the ranking, although the mean height of one tree (number 89) does reach the overall, mean. The coll e c t i o n s from Vancouver Island range from below the mean (tree 155) to well above (tree 157) and do not appear to d i f f e r i n t h e i r d i s t r i b u t i o n s from others of comparable elevation, as indicated previously. Height : diameter r a t i o Although the height means exhibit some pattern with elevation of source, the high i r r e g u l a r i t y of means from ind i v i d u a l trees suggests that height alone i s not a sensitive indicator of the char a c t e r i s t i c s of the environment at the seed source. Since tree form generally changes with a l t i t u d e , the r a t i o of height to diameter was calculated and analysed by region and elevation, as above. Plots of mean diameter for each cone parent over a l t i t u d e indicated that diameter remained the same or decreased i n s i g n i f i c a n t l y as elevation of seed parent increased. Thus, a change i n the height : diameter (H:D) r a t i o reflected mainly a decrease i n height. Analysis of variance of H:D r a t i o indicated that region, elevation i n region and parent tree within elevation were a l l important sources of var i a t i o n . Linear regressions of the height : diameter r a t i o with elevation were s i g n i f i c a n t but weaker than for height alone, r e f l e c t i n g the lack of change i n diameter with elevation. Seedlings originating from stands at si m i l a r elevations were compared. Those from Nelson at 2000 feet were r e l a t i v e l y t a l l e r at the same diameter than seedlings from Haney at 1800 feet and Seymour at 1900 feet . Their H:D r a t i o were 6.02, 4.93 and 5.10 cm. per mm. f o r Nelson, Haney and Seymour, r e s p e c t i v e l y (Table 3-10 ). The d i f f e r e n c e between the Haney and Seymour c o l l e c t i o n s was not s i g n i f i c a n t . The number of parent trees i n the provenances are: Nelson, 4; Haney, 10; Seymour, 5. The samples from the Nelson and Seymour areas are small, and therefore may be a t y p i c a l . Assuming that the parent trees i n each stand or i g i n a t e d from seed sources at the same elevation, these r e s u l t s suggest that progeny from t h i s I n t e r i o r stand possess a natural d i f f e r e n c e i n the r e l a t i o n s h i p between seedling height and diameter than the Coastal sources sampled. Brix (1971) found that western hemlock seedlings reared i n growth rooms responded strongly i n diameter to d a i l y degree hours and to l i g h t i n t e n s i t y . Height responded only weakly to l i g h t i n t e n s i t y and l e s s dramatically to l i g h t period above the optimum regime of 18 hours l i g h t - 18 hours dark. Since diameter i s the denominator i n the H:D r a t i o , a s l i g h t d i f f e r e n c e i n temperature of l i g h t i n t e n s i t y between portions of the greenhouse could have caused a large d i f f e r e n c e i n height : diameter r a t i o of the respective seedlings. Therefore, r e s u l t s from other environments should be compared before firmer con-clusions about the stockiness of I n t e r i o r vs. Coastal provenances a r e j u s t i f i e d . Perhaps the d i f f e r e n c e apparent here i s a r e f l e c t i o n of the lower exposure of the parent trees at the three s i t e s . Those from Nelson are growing i n the v a l l e y bottom, whereas the other two c o l l e c t i o n s are from high on mountainsides where they are more exposed to wind stress 66 than the Nelson source. Another possible selective source working i n a simi l a r way i s snowfall, p a r t i c u l a r l y of heavy; wet , snow. The stress that can be exerted on a tree's crown and stem by heavy snow can be con-siderable, occasionally snapping the stem. I f t h i s happens, insect or disease attack probably w i l l weaken the tree to the extent that i t i s removed from the breeding population. Thornburgh (1969) concluded that winter snows pressed down the supple young germinants of western hemlock, which then could remain buried a l l summer under natural debris. P a c i f i c s i l v e r f i r seedling (Abies amabilis (Dougl.) Forb.) did not:*suffer s i m i l a r l y because they were naturally more erect and s t i f f , allowing the snow to slough o f f . Since the strength of a cylinder i s related to the cross-sectional area per unit length,.a s l i g h t change i n stem diameter of a seedling means a much greater change i n strength. Thus, the change i n height : diameter r a t i o with a l t i t u d e probably i s under selective pressure to accommodate! the increasing snow load. F a l l s of such snow are quite common at middle elevations on the lower Coastal area of B r i t i s h Columbia. Similar snowfalls occur at low altitudes i n the Nelson area also. Perhaps a combination of reduced wind and snow stress may mean stocky stem form i s not under such strong selection pressure at lakeside near Nelson as at comparable elevations, but more exposed locations, on the lower Coast. Further comparisons were made between Coastal sources only. The seedlings from the lowest elevations at Haney and Seymour did not d i f f e r i n t h e i r H:D r a t i o s . The progeny from the Haney 500-foot, Seymour 400-foot, and Caycuse 600-foot c o l l e c t i o n s were tested for differences. Haney stock showed a higher mean height : diameter value than the others: 5.84 vs. 5.35 and 5.39 cm. per mm. respectively. The numbers of parent 67 trees tested were: Haney 12, Seymour 8, Caycuse 9, which should provide a r e l i a b l e estimate of the populations i n each area. The f i n a l comparison made was between Seymour at 1000 feet and Paterson Lake at 800 feet. Seymour produced stockier seedlings than Paterson Lake: 5:37 vs. 5.97 cm. per mm. The means for the f i v e trees per provenance represented ranged from 4.83 to 6.53. The height: diameter r a t i o s of the progeny from the highest and lowest elevations at Nelson, Haney and Seymour were tested for significance. In each case there was a s i g n i f i c a n t increase i n taper at the highest ele-vation, as indicated by the s i g n i f i c a n t l i n e a r regressions of height on elevation discussed e a r l i e r . Furthermore, there were differences between the intermediate c o l l e c t i o n s . For Nelson, the progeny from the stand at lake l e v e l were more slender than the others. Although the higher-ele-vation stands dif f e r e d by 1700 feet, t h e i r H:D r a t i o s were not diff e r e n t . For the Haney material, no differences occurred between successive ele--vations, except for the coll e c t i o n s from 1300 and 1800 feet. These both di f f e r e d from the more slender trees from the lowest elevation (100 f e e t ) . Similar r e s u l t s were obtained for co l l e c t i o n s between successive elevations on Mount Seymour: Si g n i f i c a n t differences did occur between the lowest (50 feet) source and the two highest sources (1900 and 2750 fe e t ) . No other comparison gave a s i g n i f i c a n t difference, Comparisons were made also between the progenies of individual trees within stands. Considerable v a r i a t i o n was found. As mentioned previously, the Inte r i o r (Nelson) c o l l e c t i o n from 2000 feet was more slender than two Coastal sources from a similar elevation: Haney 1800 feet and Mount Seymour 1900 feet. Yet within the Nelson stand a s i g n i f i c a n t difference occurred between the taper of the progeny of tree 12 (6.79 68 cm. per mm.) and tree 15 (5.44 cm. per mm.). From the other stands, the r a t i o ranged from 5.82 to 4.02 cm. per mm. The lowest value f o r the western hemlock c o l l e c t i o n s (3.94 cm. per mm.) came from tree 10, located at 5200 feet near Nelson. The highest value f o r mountain hemlock (0.570 cm. per mm.) came from tree 105. The d i f f e r e n c e i s s i g n i f i c a n t s t a t i s t i c a l l y . As found f o r height alone, trees g i v i n g s i m i l a r stem taper i n seedlings of t h i s s i z e could be found i n a l l stands. With v a r i a b i l i t y of t h i s kind occurring within stands, and lacking s u f f i c i e n t knowledge of the response i n height and diameter proportions by hemlock seedlings from d i f f e r e n t elevations to d i f f e r e n t environments, such as obtained for Douglas-fir by Hermann and Lavender (1968), the use of a provenance mean value of height:diameter r a t i o i s questionable, since i t does not i n d i c a t e anything c h a r a c t e r i s t i c of d i f f e r e n t areas. The height:diameter r a t i o s of western and mountain hemlocks appear to be d i f f e r e n t at t h i s stage, but there are more p o s i t i v e ways to d i s t i n g u i s h the species. Branch measures The v a r i a b l e s maximum branch length and number of branches were examined to t e s t t h e i r patterns of v a r i a t i o n between and within species and populations r e l a t i v e to height and H:D r a t i o . Maximum branch length From Table 3-10 i t i s c l e a r that maximum branch length i n western hemlock drops as seed source elevation increases, but i r r e g u l a r l y . In f a c t , the greatest lengths f o r Nelson and Seymour were from intermediate elevations: 3500 fee t and 1000 feet r e s p e c t i v e l y . Although the number of surviving trees i n a family ranged between 13 and 48, t h i s v a r i a t i o n d i d not a f f e c t branch length c o n s i s t e n t l y . Closest c o r r e l a t i o n between branch length and any other variable occurred with a diameter. Elevation, height and number of branches usually added nothing s i g n i f i c a n t to the regression. However, as diameter changed least with elevation and height most, regressions of branch length on height were made for each area. Since a s i g n i f i c a n t , p o s i t i v e relationship was found, the differences i n height were removed by covariance analyses. Tests were then conducted on the adjusted means. The o r i g i n a l and the adjusted means of branch length are presented by provenance i n Table 3-12. Ba r t l e t t ' s test of the variances of longest branch with elevation for the Nelson, Haney and Seymour coll e c t i o n s found the variances non-homogeneous, so that, as before, conclusions based on analyses of these data are not f u l l y supportable. ANOVA of the transformed data indicated that both elevation and parent trees within elevation were s i g n i f i c a n t sources of va r i a t i o n with-i n the Nelson, Haney and Seymour material, and that parent tree was im-portant for the Caycuse and Paterson Lake progeny. The means i n Table 3-12 show.a regular trend with elevation only for the Haney c o l l e c t i o n , decreas-ing from 6.68 cm. at 100 feet to 5.55 cm. at 1800 feet. For both the Nelson and Seymour coll e c t i o n s an e r r a t i c trend i s apparent, since the maximum . length occurred at an intermediate elevation. This i s most pronounced for the Nelson material where the mean for the 3500-foot source i s 6.83 cm., and that for the 200-foot source i s only 6.11 cm. Over a similar span on Mount Seymour the mean decreased s l i g h t l y , but strongly at Haney. For the Mount Seymour material only, the c o l l e c t i o n from 400 feet gave a mean value lower than that from 1000 feet and equal to that from 1900 feet. Table 3-12 Actual and adjusted means of longest branch i n cm. a f t e r covariance removal of mean height d i f f e r e n c e s . Western hemlock only. Area Status E l e v a t i o n of c o l l e c t i o n (feet)  Nelson 2000 3500 . 5200 Actual 6.21 6.86 4.88 Adjusted 6.11 6.83 5.57 Haney 100 500 1300 1800 Actual 7.07 6.09 5.71 5.34 Adjusted 6.68 6.08 5.64 5.55 Seymour 50 400 1000 1900 2750 Actual 6.37 5.76 7.10 5.94 5.33 Adjusted 6.11 5.87 7.07 5.96 5.55 Caycuse 600  Actual 6.62 Adjusted 6.18 Paterson 800 Actual 6.32 Adjusted 5.92 Linear regressions of the form: Branch length = a + b (elevation) were calculated on the adjusted branch lengths for the Nelson, Haney and Seymour mat e r i a l . They are presented i n Table 3-13. 71 Table 3-13 Linear regressions of adjusted branch length with change i n elevation for Nelson, Haney and Seymour col l e c t i o n s . Western hemlock only. Area Elevation Regression Summary Limits Intercept Slope Standard r S i g n i f . (feet) (cm) (cm./lOOO f t . ) error  Nelson 2000-5200 -0.347 0.102 2.081 .004 N.S. Haney 100-1800 0.493 -0.524 1.728 .036 Seymour 50-2750 0.194 -0.194 1.887 .008 N.S. "*"*** Si g n i f i c a n t at the 0.1 percent l e v e l Only the Haney material produced a s i g n i f i c a n t trend - branch length decreased s l i g h t l y as elevation of c o l l e c t i o n increased, as noted before. However, only 3.65 percent of the v a r i a t i o n i n branch length was explained by elevation of o r i g i n , whereas a very high proportion of the v a r i a t i o n was attributable to cone parent within elevations. The values for co l l e c t i o n s from sim i l a r elevations were compared s t a t i s t i c a l l y . Mean branch length from the Caycuse c o l l e c t i o n was similar to those from comparable elevations on the Lower Mainland (Haney and Seymour). The Paterson Lake material gave a mean smaller than the Seymour 1000:-foot material (5.92 vs. 7.07 cm.). Since t h i s c o l l e c t i o n from Mount Seymour gave the longest branches of a l l from that area, the meaning of t h i s difference i s not clear. The Nelson c o l l e c t i o n from 2000 feet produced longer branches than that from Haney at 1800 feet, but about equal to the Seymour material from 1900 feet. The low-elevation c o l l e c t i o n from Haney gave longer branches than i t s counterpart from the base of Mount Seymour, but t h i s had reversed for the c o l l e c t i o n s from 1300 and 1000 feet. 72 As f o r height, and height:diameter r a t i o , parent tree was very highly s i g n i f i c a n t as a source of v a r i a t i o n i n the ANOVA of longest branch. The greatest d i f f e r e n c e within a stand occurred i n the Paterson Lake c o l l e c t i o n . Tree 163 had a mean value of 4.83 cm., and tree 161 of 8.23 cm. A s i m i l a r d i f f e r e n c e was found between trees 10 and 9 from 5200 feet a t Nelson. Judging by the standard deviations, longest branch length varied l e s s at high than at low elevations. But when the c o e f f i c i e n t of v a r i a t i o n was calcul a t e d , no trend with increased elevation was apparent. In f a c t , the Nelson material from 5200 feet was sharply more v a r i a b l e than that below i t ; the c o e f f i c i e n t s of v a r i a t i o n were 32.4%, 33.1% and 58.8% for the c o l l e c t i o n s from 2000, 3500 and 5200 feet, r e s p e c t i v e l y . This change was due l a r g e l y to the presence of tree 10 i n the highest c o l l e c t i o n . I t produced highly v a r i a b l e heights i n both t h i s t r i a l and i n the bare-root nursery t r i a l . Examination of an X-ray p l a t e of other samples of t h i s seed showed many with incompletely developed "endosperm" and embryos. Probably, t h i s i s i n part a consequence of c o l l e c t i n g the seed before they were f u l l y matured, and t h i s immaturity has c a r r i e d over as low vigour of the progeny. Maximum branch length was measured with the hope that i t could add useful information about the progeny from various stands to that obtainable using H, D, and H T D r a t i o . A s i g n i f i c a n t and•regular reduction i n the mean longest branch, a f t e r adjustment f o r di f f e r e n c e s i n mean height, occurs with an increase i n elevation only i n the Haney progeny. Height decreased more e r r a t i c a l l y with elevation there, so longest branch i s not r i g i d l y determined by height. Since no such trend was found for Nelson and Seymour material when height differences had been removed, they allow a dif f e r e n t conclusion. Although longest branch was not affected by the number of surviving trees i n a block at t h i s stage, i t l i k e l y would be i f they were kept much longer i n these f l a t s . The strength of the s i g n i f i -cance of cone parent i n the v a r i a t i o n of t h i s feature indicates that i t could be a useful feature of early progeny test evaluation for fami l i e s , but not for provenances i f they do not contain more broadly-based sources than t h i s study. Number of branches ANOVA of number of branches for western hemlock seedlings revealed that i t varied s i g n i f i c a n t l y with seedling height. Before further analysis was attempted the number of branches was adjusted by covariance for d i f f e r -ences i n seedling height. Further ANOVA using these adjusted values i n d i -cated strong v a r i a t i o n between families and weaker va r i a t i o n between eleva-tions and areas. The highest number of branches occurred among: low-eleva-t i o n progeny. Regression values for the trends of adjusted branch numbers with elevation for the Nelson, Haney and Seymour areas appear i n Table 3-14. Table 3-14 Linear regressions of adjusted number of branches of western hemlock families with elevation of c o l l e c t i o n : Nelson, Haney and Seymour c o l l e c t i o n s . Number of Change pr. Area Families Seedlings 1000 feet Intercept r Si g n i f . Nelson 11 162 -0. .633 2. .431 0. .210 ** Haney 43 430 -0. .679 1. .474 0. .071 ** Seymour 40 388 -0. .314 1, .045 0. .082 ** A highly s i g n i f i c a n t relationship was found i n each case. The slope of the Nelson and Haney material was s i g n i f i c a n t l y steeper than for the Seymour c o l l e c t i o n , prohibiting pooling a l l data into a common re-gression. The strongest association with a l t i t u d e was found for the Nelson 2 progenies (r = 0.210). This c o l l e c t i o n covered the broadest a l t i t u d i n a l range, permitting clearest expression of the response i n branch number. The low coe f f i c i e n t s of corr e l a t i o n , and inspection of the plots of branch number with a l t i t u d e , indicate that most of the v a r i a b i l i t y i n t h i s feature i s associated with family, not elevation. This was reflected i n the strength of the "family" term i n the ANOVA. Since a negative l i n e a r relationship exists between adjusted branch number and elevation, probably the reduction found i s a r e f l e c t i o n of response to environmental pressure. A high branch number provides more surface to support f o l i a g e , increasing the chance that a seedling with a high branch number w i l l be able to outgrow i t s competitors. This advantage w i l l be counteracted by snow pressure at high a l t i t u d e s , increasing the chance that branchy seedlings w i l l be bent or broken (Thornburgh, 1969). This measure followed the trend that emerged for seedling height : decrease with a l t i t u d e of seed source. This trend remained when height differences between seedlings were removed, indicating that height and number of branches are separable to some degree, and that selection for reduced number of branches need not be at a s a c r i f i c e of seedling height. S i m i l a r l y , because of the intra-population v a r i a b i l i t y i n t h i s feature due to cone parent, selection for reduced branch number may proceed without loss of adaptation to l o c a l climate. 75 Stronger regressions with a l t i t u d e were found f o r branch number than f o r branch length; each of the areas tested gave a s i g n i f i c a n t r e l a t i o n s h i p f o r branch number, whereas branch length was r e l a t e d s i g n i f i c a n t l y with a l t i t u d e only f o r Haney. This makes branch number a more useful parameter than branch length, but the high intra-provenance v a r i a b i l i t y i n branch number associated with seed parent means that t h i s feature i s not a s e n s i t i v e i n d i c a t o r of population response to the environment. For that reason, further analysis was not conducted . Frost damage A l l seedlings i n each family were appraised f o r f r o s t damage at the time of measuring: seedlings were t a l l i e d "healthy" or "frost-damaged". Any seedlings with brown leaves retained to the t i p or with brown leaves retained below a bare t i p were considered frost-damaged. (Minimum screen temperature at the Haney Forest Administration s t a t i o n beside the con-tain e r compound di d not f a l l to f r e e z i n g u n t i l Nov. 21 and 22nd., when i t plunged sharply to 20°F, a drop of 13 degrees from the former minimum. Grass minimum could have been s l i g h t l y lower. Probably the f r o s t damage occurred during these n i g h t s ) . The r e s u l t s are summarized by family and provenance i n Table 3-15. No damage was found on the mountain hemlock f a m i l i e s . In t h i s respect they are s i m i l a r to;the western hemlock population c o l l e c t e d nearest: at 2750 feet on Mount Seymour, but d i f f e r e n t from the Nelson 5200 feet c o l l e c t i o n . Some f r o s t damage occurred i n a l l but two of the western hemlock 76 Table 3-15 Percentage of plug-grown seedlings showing f r o s t damage by area and elevation of o r i g i n . Area Elevation No. of No. of Percent frost-and species (feet) families seedlings damaged Western hemlock Nelson 2000 3500 5200 4 4 3 128 136 82 0.0 6.61 3.65 Haney 100 500 1300 1800 9 12 12 10 297 385 414 353 11.44 2.59 3.38 2.29 Mount Seymour 50 400 1000 1900 2750 9 8 5 5 7 245 268 149 127 196 20.40 13.05 4.02 1.38 0.0 Caycuse 600 243 10.28 Paterson Lake 800 158 1.26 B, Mountain hemlock Mount Seymour 3200 94 0.0 provenances. Those western hemlock provenances without damage were from the lowest a l t i t u d e at Nelson (2000 feet) and from the highest a l t i t u d e (2750 feet) on Mount Seymour. Otherwise, there was a general trend of decreasing damage with increasing elevation of seed source. This was especially so with the c o l l e c t i o n from Mount Seymour, where fr o s t damage decreased from 20.4% at 50 feet to n i l at 2750 feet. The trend was not as clear for Haney material, where the provenance from 100 feet was damaged 11.44%, but those from 500, 1300 and 1800 feet were about equally damaged. Any similar trend from the Nelson collections was confused by the absence of fr o s t damage i n the low-elevation c o l l e c t i o n , but damage i n the high-elevation provenance was Tess than that i n the middle-altitude provenance (3.6% vs. 6.6%). Inter-family v a r i a t i o n i n f r o s t damage was very pronounced i n the high-a l t i t u d e populations, p a r t i c u l a r l y those from Nelson just discussed, where a single tree i n each stand produced a l l the damaged seedlings entered i n the provenance average. With a maximum of four parent trees from which to obtain data, there i s a f a i r l y high p r o b a b i l i t y that the parent trees chosen do not represent the l o c a l populations accurately, and that a l l of'the Nelson provenances are equally susceptible to f r o s t . The same might be true for the progeny from 1900 feet on Mount Seymour, where progeny of only one tree (number 61) from a 5-tree sample sustained damage. But the ten-tree c o l l e c -t i o n from a comparable elevation (1800 feet) at Haney exhibited about the same degree of f r o s t damage from two trees, numbers 91 and 93. Assuming that t h i s c o l l e c t i o n produced a r e l i a b l e response f o r . t h i s elevation and lati t u d e i n the lower Coastal region, the results from the Seymour c o l l e c t i o n seem comparable. Since only two of 144 seedlings i n the l a t t e r provenance were damaged, a single additional damaged tree would have raised the 78 provenance value to 2.1%, close to that from the Haney c o l l e c t i o n . . The heaviest damage (81.2%) occurred among the progeny of tree 71, growing at about 50 feet elevation at the foot of Mount Seymour. But t h i s value was based on only 16 surviving t r e e s , so chance alone might account f or i t being that high. Heavy damage (65.8% of 38 trees) was found among the progeny of tree 77 from 400 fee t on Mount Seymour. Largely because of that s i n g l e cone parent, the incidence of f r o s t damage was higher i n that provenance than i n the 500-foot provenance from Haney (2.6%). In general, the frequency and percentage of parent trees e x h i b i t i n g damage decreased with increasing a l t i t u d e . Inspection of the damage frequencies revealed no c o r r e l a t i o n between them and the transplanting of seedlings the previous summer. The Haney C o l l e c t i o n s were based on s u f f i c i e n t trees at each elevation to permit comparison of r e s u l t s . They are: 100 fee t elevation, 7 of 9 trees (77.7%); 500 fee t elevation, 4 of 12 trees (33%); 1300 feet elevation, 7 of 12 trees (58.2%); 1800 feet elevation, 2 of 10 trees (20%). Evidently the pattern i s not regular, but with c o l l e c t i o n s of 12 trees per provenance sampling error could account f o r both part of the apparent trend of decreased incidence of f r o s t damage with increased a l t i t u d e and the apparent anomaly of higher damage i n the 1300-foot provenance than i n the 500-foot provenance. A s i m i l a r s i t u a t i o n e x i s t s i n the Mount Seymour provenances, where the 5-tree c o l l e c t i o n from 1000 feet exhibited a higher frequency of damage than the c o l l e c t i o n from 400 fee t (60% vs. 37.4%), although only three parent trees contributed damaged seedlings i n each case. 79 Generally, the pattern of fr o s t damage i s p a r a l l e l to that of population mean height (Table 3-10 )• However, the relationship i s not clear-cut, since mean height for the Nelson population decreased with elevation, but damage did not, and the reversal of mean height between 400 and 1000 feet on Mount Seymour does not correspond with the steady decline i n fr o s t damage. Inconsistencies are apparent when populations of sim i l a r mean height are compared. (Tables 3-10and 3-15 ). As examples, Caycuse and Paterson Lake d i f f e r i n f r o s t damage, whereas Haney 1300 feet and Mount Seymour 1000 feet are quite si m i l a r i n both height and f r o s t damage. No further insights are obtainable from these r e s u l t s ; those from the bare-root nursery study (Section 3.52) are better founded, with r e p l i c a t i o n and another mountain hemlock family to study. .3.52 Bare-root nursery study A nursery t r i a l was established i n the Forestry nursery, U.B.C. south campus, i n the spring of 1970 to provide information on the following points by species, population and seed parent: rate and completeness of germination of s t r a t i f i e d vs. u n s t r a t i f i e d seed; the frequency and nature of any morphologically-aberrant types; su r v i v a l ; bud set and bud burst patterns; and the occurrence, type and severity of any damage caused by the climate. 80 Materials and methods Four r e p l i c a t e s , each including rows of 100 seeds per treatment ( s t r a t i f i e d and unstratified) were established for 88 cone parents. For three other trees, numbers 61, 62 and 78, a l l western hemlocks, shortage of seed required elimination of the u n s t r a t i f i e d portion of the t r i a l . An error i n counting occurred i n one sample from tree 77, a western hemlock - re s u l t s from re p l i c a t e 2 s t r a t i f i e d are based on only 64 seeds (counted at the time of sowing). The 100-seed samples were withdrawn i n two l o t s of 50. Only complete seeds were counted. The remaining seeds were well mixed before any further samples were obtained. A l l seeds were returned to the cold o o room (1 +^ 1 C) u n t i l a l l l o t s were counted and ready for soaking. Seed samples were soaked and s t r a t i f i e d for 21 days as described i n Section 3.51. Seeds for each r e p l i c a t e were put to soak on successive days to maintain a constant c h i l l i n g period. The nursery beds were framed i n rough cedar planking and f i l l e d f i r s t with the available s o i l material remaining after vandals had destroyed the nursery the previous December. I t was a sand-peat mixture. Over t h i s , Baker's (1957) "Cal. C" mix was added to f i l l the bed to approximately 8 inches. This was l e v e l l e d , raked and pressed by hand. Seeds were sown i n a row across the bed. Rows were 4 inches apart i n i t i a l l y , but those formed for the s t r a t i f i e d seed were placed between, making the f i n a l spacing 2 inches between rows. Seed was not sown within two inches of the sideboards. Spacing of the seed was only approximately equal within the row. Fine g r a n i t i c sand was poured over the seed to protect i t from drying when a row was sown. The order of parent trees was completely random within and between r e p l i c a t e s . Each row was i d e n t i f i e d by a tag at the margin of the bed, carrying r e p l i c a t e number, tree number and a l e t t e r to i d e n t i f y the rows of u n s t r a t i f i e d seed (Figure 3-13) . Each r e p l i c a t e was watered down and shaded as soon a f t e r sowing as p o s s i b l e . Shade was provided by moveable screens 4 feet square con-s i s t i n g of f i n e f i b r e g l a s s (13 mesh., per inch) screening stapled to a wooden frame of s i x - i n c h lumber. They were moved l a t e r a l l y approximately weekly so that a l l areas of the bed were for a time under the f u l l shade of the overlapping frames. Water was applied by powerful r o t a t i n g s p r i n k l e r s d a i l y u n t i l mid August, and l e s s frequently a f t e r that. However, t h i s s p r i n k l e r system was not i n s t a l l e d u n t i l 3 weeks a f t e r sowing. In the meantime, water was applied by hand. This meant les s even watering than possible using the s p r i n k l e r s , and may have caused i r r e g u l a r i t y i n the germination and early development of the trees. F e r t i l i z e r was applied i n s o l u t i o n on the following dates: July 5 to r e p l i c a t e 1 (to t e s t the response of the seedlings for possible damage from the f e r t i l i z e r i o n s ) , July 14 to r e p l i c a t e s 2- 4. "Hi-Sol" a "balanced" commercial f e r t i l i z e r was measured to apply 20 pounds per acre of nitrogen, as recommended by van den Driessche (1969a). The amount of each r e p l i c a t e was divided into four equal B e d Rep. 2 Rep. 1 Rep. 4 i i | unused bed I 1 r- -\ Scale: 1 cm Rep. 2 j Rep. 3 I I 1 IR! 4 i i 1 i 1 i / / / \ \ \ S t r a t i f i e d Family no's. / / ae ms oo ez 4-o se \ \ i i 3» at SO 29 z sz so U n s t r a t i f i e d = 6 . 2 5 f t . Scale: 1 cm = 6 i n . Figure 3 - 1 3 Plan of beds i n nursery study, U.B.C. Forestry nursery C D 83 parts that were dissolved into one gallon of water. This was applied by hand to the appropriate length of bed, then water was sprinkled l i g h t l y over the bed to wash the f e r t i l i z e r from the seedlings. The beds were checked regularly to detect the f i r s t germinants. Counts were made weekly on the same day as the seed had been sown (Wednesday, Thursday, etc.) for seven weeks. A seedling was counted when the seedcoat, borne upward by the hypocotyl, was clear of the ground, or when the cotyledons were free of the ground. Coloured p l a s t i c rings were placed over each seedling as i t was counted. The week of counting was indicated by different-coloured rings. Damaged trees, presumably caused by a ground-dwelling "cutworm", were f i r s t noticed on June 15, 1970. Cotyledons had been cut back or removed completely, or seedlings were cut o f f , only a stub remaining i n the r i n g . This damage occurred i n d e f i n i t e l o c a l i t i e s , often being limited to 2 or 3 adjacent rows. Occasionally nearly half a row was affected. "Diazinon" (0, O-diethyl-0 (2-isopropyl-4 methyl-6-pyrimidyl) phosphorothioate) was applied i n solution of recommended strength (10.5 cc per gallon) to a l l beds on June 17. The beds were checked for more damage the next two days but none was found. On July 1, 2 and 3 more damage of the same kind was found. Diazinon was applied again at the same strength. When more damage and two larvae, apparently of a cutting insect, were found on July 9, a Malathion drench was applied to the entire experiment that evening at a dosage of 1 teaspoon per 2 gallons of water. No more "cutworm" damage was noted after that. 84 Some s l i g h t b i r d damage occurred. Some epicotyls were damaged or seedlings pulled out by birds apparently attracted by insects on the plants or the coloured rings around the tree bases. A l l t h i s damage was noted and removed before survival values were computed. Shading was reduced i n early September by removing the screens, except during warm days. They were removed for- the winter on September 16th. In May of 1971 "Hi-Sol" f e r t i l i z e r was applied at a rate of 40 l b . of nitrogen per acre by spreading a measured amount along a s l i t between alternate rows, then digging i t i n s l i g h t l y . This was an attempt to locate the f e r t i l i z e r i n a position from which i t could be withdrawn readily by the plant. Shades were placed over the seedbeds i n A p r i l , 1971, when bud burst had begun. They were kept on u n t i l September 3, 1971. As i n 1970 they were moved f a i r l y regularly to d i s t r i b u t e the heavy shade under the screen frames as evenly as possible over the seedbed area. The screens were raised above the sideboards during the season to prevent the screens pressing down the leaders of the t a l l e r trees. This l e t d i r e c t sunlight shine on seedlings along the west side of the beds, causing some "bleaching" of trees there. Sunscald damage occurred to trees i n a l l replicates when the shades were removed completely after several days of wet, cloudy weather i n June. Due to an oversight, the shade frames i n re p l i c a t e 4, Bed 3 had not been shifted for about 10 days. The seedlings were exposed to about half a day of bright sun, although the day was cool. Damage to the trees was evident the next morning. Bands of trees across the bed were bleached pale. The bands corresponded to the most recent po s i t i o n of the overlapped screen frames. Where the alignment of the~rows and the 85 screens did not coincide, trees i n only part of the rows were affected, and a few trees were found damaged on one side only. The trees had been f u l l y i n an actively-growing state when exposed. Some trees i n rep l i c a t e 4 l o s t leaves during the next few weeks and were much smaller than t h e i r neighbours by the end of the season. No damage of the kind or extent presumably due to cutworms the year before was found during 1971. Only a few cases of chewed foliage or trimmed epicotyls were discovered. Counts of the number of trees surviving i n each row were made at the end of each growing season. In 1971 t h i s included the number of weak, damaged and "odd" trees; ones with noticeable abnormalities of si z e , colour or form. Results Germination Western hemlock Mean values of t o t a l germination percent of seeds sown, germinative energy and percent of f i n a l germination represented by the point of c u l -mination of 'germinative energy appear i n Table 3-16. Table 3-16 Values of nursery germination for s t r a t i f i e d and un-s t r a t i f i e d western hemlock seed Treatment Total Germinative Fraction of germination energy ' (days) f i n a l % % S t r a t i f i e d 58.56 Uns t r a t i f i e d 62.08 Difference 3.52 Significance * 38.91 41.12 2.21 84.48 86.12 1.64 N.S. 86 T o t a l germination was assessed for a l l trees for which there was equal representation of s t r a t i f i e d and u n s t r a t i f i e d rows. T o t a l germination was depressed s l i g h t l y but s i g n i f i c a n t l y by s t r a t i f i c a t i o n : mean germination of s t r a t i f i e d seed was 58.6% compared to 62.1% for u n s t r a t i f i e d seed. Edwards (1973) found s l i g h t l y higher germinative capacity following s t r a t i f i c a t i o n . The data were examined by parent tree (Figure 3-14). Some c o l l e c t i o n s gave s l i g h t l y higher germination a f t e r s t r a t i f i c a t i o n . These did not follow any pattern ass o c i a t i o n with c o l l e c t i o n date, a l t i t u d e , area of o r i g i n or t o t a l germination; higher germination by s t r a t i f i e d seed occurred where t o t a l germination of seed sown was high or low. Inspection of values from i n d i v i d u a l trees showing sharp d i f f e r e n c e s between s t r a t i f i e d and un-s t r a t i f i e d values revealed none with s i g n i f i c a n t d i f f e r e n c e s between treatments. In view of the marked intra-population differences i n seed maturity found by A l l e n (1958a) and evident here, perhaps s t r a t i f i c a t i o n r e s u l t e d i n the f a i l u r e of the least-developed seeds p r i o r to emergence whereas the u n s t r a t i f i e d seeds were not subjected to t h i s stress, and more seeds survived to germinate. C a l c u l a t i o n of germinative energy was conducted on the data from 22 trees chosen f o r t h e i r high proportion of f i l l e d seed. Germinative energy was c a l c u l a t e d as follows: cumulative values of seedlings emerged were p l o t t e d by computer on rectangular co-ordinates; a smooth freehand curve was drawn through these points (Figure 3-15); a s t r a i g h t l i n e was drawn from the date of sowing to touch the curve at i t s highest point permitting a tangent; the date corresponding to t h i s point was determined, as well as the predicted value of seeds germinated f o r the ioo h 90 80 70 60 50 40 30 20 h . Legend: 0 Nelson O Haney Seymour + Caycuse A . P a t e r s o n Lc.ke / ° °o/<P $> o 164 V O o • 0 •/ o o 4 o 6? loo A?' 0 T 92 V 161 _L JL 10 90 100 20 30 40 50 60 70 80 U n s t r a t i f i e d g e r m i n a t i o n p e r c e n t F i g u r e 3-14 Comparison o f mean t o t a l g e r mination f o r s t r a t i f i e d and u n s t r a t i f i e d western hemlock seed. oo 88 100 I— 90 T3 CD CP rt CU e cn •H i r t T>-Q) Q) cn CH o rt s C CL) > •rt -P 3 O 80 v o l 60 |— 50 40 30 20 10 L S t r a t i f i e d o—o-U n s t r a t i f i e d • • Fracti o n of f i n a l S t r a t i f i e d : 77 - 89 = 0.865 U n s t r a t i f i e d : 63 - 73 = 0.863 Germinative energy: S t r a t i f i e d : 31 days U n s t r a t i f i e d : 37 days 0 0 1 7 2 14 3 21 4 28 ." 5 .35 6 42 7 8 9 49 56 63 weeks 16 Days 102 T i m e a f t e r s o w i n g Figure 3-15 Course of germination f o r s t r a t i f i e d and u n s t r a t i f i e d mountain hemlock seed, and d e r i v a t i o n of germinative energy and f r a c -t i o n of f i n a l . Tree 105, Replicate 3. 89 same date. The number of days elapsed between sowing and the date determined by the tangent point was the "germinative energy" value; the predicted value of seeds germinated on that date formed the "fraction of f i n a l " value when divided by the t o t a l germinated seeds i n the row. ANOVA of germinative energy showed s t r a t i f i c a t i o n had hastened germination, as found also by Edwards (1973). However, the gain i n time was small, only 2.2 days over the u n s t r a t i f i e d time of 41.1 days. This i s a very s l i g h t decrease i n incubation time when the elapsed time i s nearly seven weeks. Edwards (1973) found a.decrease i n R^ -Q of nearly 5 days following s t r a t i f i c a t i o n and incubation i n Jacobsen apparatus. Including the time required to prepare the seeds for s t r a t i f i c a t i o n , which was one-half day per r e p l i c a t e , s t r a t i f i c a t i o n represents no saving i n time expended. There could be other merit i n s t r a t i f i c a t i o n on an operational basis, such as establishment of a more even stand, or emergence over a shorter portion of the elapsed time (Edwards, 1973), but no such benefits appeared here. Perhaps the usual germinative acceleration from s t r a t i f i c a t i o n was masked, here by the i n a b i l i t y to supply water regularly for the two weeks following sowing, when the sprinkling system was not i n s t a l l e d , or the less precise conditions outdoors versus the laboratory. From Figure 3~i4, intra-provenance v a r i a b i l i t y was much stronger than inter-provenance v a r i a t i o n i n response to s t r a t i f i c a t i o n . This probably i s associated most with difference i n seed maturity, the less mature co l l e c t i o n s being affected adversely by s t r a t i f i c a t i o n . Consider-able v a r i a t i o n i n cone colour between green and brown was apparent between trees i n the stands during c o l l e c t i o n , indicating that s i m i l a r v a r i a t i o n i n seed maturity could occur between trees, as found by Alle n (1958a). In addition to t h i s source of var i a t i o n must be added va r i a t i o n i n f i l l e d seed percent. Inadequate checking on the X-ray flourescent screen during seed sorting gave poor separation of f i l l e d and empty seed for some co l l e c t i o n s . Consequently, the germinative capacity varied con-siderably, and the spread of t o t a l germination seen i s not due only to immature seed. As shown i n Table 3-16, there was no s i g n i f i c a n t difference be-tween s t r a t i f i e d and u n s t r a t i f i e d rows i n the proportion of f i n a l germin-ation reached at the "germinative energy" point. Their values were 84.5% and 86.1%, respectively. This means that, i n general, the curves of cumulative germination were shaped the same beyond the point defining germinative energy. Consequently, the main difference between the ger-minative behaviour of the s t r a t i f i e d and u n s t r a t i f i e d rows lay i n the speed with which the culmination of germination was reached. As the difference i n germinative speed was so s l i g h t , and because the deter*-mination of germinative energy was so time-consuming, i t was not pursued further i n the nursery material. More sensitive evaluation of the germinative behaviour of western hemlock was conducted on seed from con-t r o l l e d crosses (Chapter 4). Mountain hemlock Mean values for the same features of germinative behaviour as discussed for western hemlock appear i n Table 3-17, arranged by seed parent. 91 Table 3-17 Values of nursery germination for s t r a t i f i e d and u n s t r a t i f i e d mountain hemlock seed. Feature S e e d P a r e n t Mean Signif. 105 106 109 Str. Uns. Str. Uns. Str. Uns. Str. Uns.  Total germ.% 81.0 81.0 86.0 90-.0 95.5 94.6 87.5 88.6 N.S. Germ. energy 29.7 34.7 30.9 36.2 32.0 33.9 30.9 34.9 * * Gays) Fraction of f i n a l 0.87 0.76 0.89 0.40 0.85 0.89 0.87 0.85 N.S. Unlike western hemlock, mountain hemlock did not germinate less following s t r a t i f i c a t i o n . Greater differences occurred between seed parents than between treatments. Similar to western hemlock, germinative energy was increased by s t r a t i f i c a t i o n , although here i t was by an average of 4 days: 30.9 vs. 34.9 days. This indicates a requirement by mountain hemlock for deeper c h i l l i n g to stimulate germination than exists i n western hemlock, and l i k e l y r e f l e c t s adaption to a colder, snowier environment than occupied by western hemlock. Rapid germination, as found for u n s t r a t i f i e d mountain hemlock seeds here i n comparison to western hemlock, would be of high selective advantage under these circumstances to permit f u l l u t i l i z a t i o n of the growing season. S t r a t i f i c a t i o n did not change appreciably the value of the fr a c t i o n of f i n a l germination reached at the germinative energy point. The mean values for s t r a t i f i e d and u n s t r a t i f i e d seed were 0.87 and 0.85, respectively. Thus, as for western hemlock, only germinative energy was increased by s t r a t i f i c a t i o n . Seed parent (family) influence on t o t a l germination appeared important. The lower value for tree 106 s t r a t i f i e d was due mostly to a sharp reduction i n r e p l i c a t e 2. A d i f f e r e n t pattern appeared for germinative energy, where the rate was not increased for tree 109. This same family was increased least i n i t s f r a c t i o n - o f - f i n a l value. With only three seed parents available for study, estimating the degree of intra-provenance v a r i a b i l i t y i n germinative behaviour by mountain hemlock i s not j u s t i f i e d . Further investigation of the germinative trends of western hemlock did not seem j u s t i f i e d i n view of the work involved to calculate germin-ative energy. Each species emerged faster following s t r a t i f i c a t i o n , but mountain hemlock was accelerated more. Whether or not s t r a t i f i e d , the two species exhibit d i f f e r e n t germinative energy i n these conditions: mountain hemlock emerged faster. This agrees with my experience with provenances of these species i n laboratory t e s t s , and suggests strong differences i n physiology between the two species i n the processes i n -volved i n seed germination. Mutants Note was made of a l l aberrant types of trees as they appeared. They were entered under one of the following classes: Albinos and semi-albinos, "weak" seedlings, and healthy "mutants". Those i n the albino categories were c l a s s i f i e d according to the classes of seedling and plant mutants of Eiche (1955). The mutant types were noted during the weekly counts and followed as long as possible. Results were analysed to determine differences 93 between species, population and parent tree. Albinos and semi-albinos The occurrence of seedlings i n t h i s class i s summarized i n Table 3-18. Despite being represented by a high number of seedlings, 2122 from 2400 seeds sown, mountain hemlock produced no chlorophyll-deficient seedlings. This may be due to the r e s t r i c t e d sample, only 3 parent trees from one population, or to a lack of the necessary genes i n the population or the species. Fourteen of the 88 western hemlock trees studied (15.9%) produced one or more chlorophyll-deficient seedlings. Overall, 20 such "mutant" seedlings occurred, constituting 0.27% of the progeny of the mutant-bearing parents, and 0.053% of the 37,815 western hemlock seedlings i n the entire nursery study. In comparison, Eiche (1955) found an o v e r a l l average of 0.152% mutants among 1,769,000 seedlings from 1016 Scots pine trees from 43 stands. Within individual f a m i l i e s , the frequency of mutant western hemlock seedlings varied between 0 and 0.9%. The range i n Eiche's material was from 0 to 25%. 94 Table 3-18 Frequency and type of f u l l y or p a r t l y c h l o r o p h y l l - d e f i c i e n t seedlings i n the nursery study by o r i g i n and parent tree. Area Elev. Parent Emer- Frequency of Chlor- Type (Feet) Tree ged o p h y l l - d e f i c i e n t (Eiche, No. seed- Type 1955) l i n g s Lethal Non-lethal No. pet No. pet. Western hemlock Nelson Haney Seymour Caycuse 2000 3500 5200 100 500 1300 1800 50 400 1000 1900 2750 600 14 18 29 37 41 51 68 70 75 77 81 57 102 491 572 569 475 622 640 399 593 557 600 639 334 379 1 N i l N i l 1 0 0 1 2 N i l 1 2 0 0 0 N i l 1 1 N i l 0.204 0.175 0.161 0.312 0.251 0.337 0.299 1 1 0 0 N i l 0 0 3 1 1 N i l 2 N i l N i l 0.176 0.210 0.539 0.167 0.156 0.598 0.264 0 N i l a l b i n a albina v i r i d i s v i r i d i s a l b i n a a l b i n a a l b i n a a l b i n a v i r i d i s xantha v i r i d i s a l b ina v i r i d i s a l b i n a Paterson Lake 800 164 421 0.238 albina Mountain Hemlock 3300 2122 N i l N i l The frequency and percentage of chlorophyll-deficient mutant types by population sampled appear i n Table 3-19. Table 3-19 Freguency and percentage of chlorophyll-deficient mutant seedlings i n hemlock populations studied i n the nursery. 1970. Area Elev. (feet) No. of parents No. of seedlings No. of mutants Mutant percent Western hemlock Nelson 2000 3500 5200 1170 840 136 1 0 0 0.085 Haney 100 500 1300 1800 8 11 11 10 4079 4729 5596 3421 0.024 0.042 0.0535 Seymour 50 400 1000 1900 2750 9 8 3 5 5 3711 4555 1410 1717 2252 3 5 0 3 1 0.081 0.110 0.175 0.044 Caycuse 600 3057 Paterson Lake 800 1152 0.087 Mountain hemlock Mt. Seymour 3300 2122 Nine of the f i f t e e n populations studied produced at least one chlorophyll-deficient seedling. The population producing the most mutants was that at 400 feet on Mount Seymour. Three of the f i v e seedlings noted there came from tree 75 (Table 3-18 ). Three other parents produced two mutants i n one class; tree 51 from 1300 feet at Haney, tree 57 from 1900 feet on Mount Seymour, and number 70 from 50 feet on Mount Seymour. 96 Some of the populations that produced no c h l o r o p h y l l - d e f i c i e n t types were represented by only a few parent trees. For instance, the Nelson c o l l e c t i o n from 5200 fee t contained only two f a m i l i e s , and that from 1000 f e e t on Mount Seymour three f a m i l i e s . But there were also c o l l e c t i o n s based on more than f i v e parents included i n t h i s category. For instance, the Haney 1800-foot population was represented by ten parents, and the Caycuse l o c a l i t y by s i x . Since the mutant from Nelson came from a 3-tree c o l l e c t i o n , i t does not seem that r e s t r i c t e d sample si z e alone i s responsible f o r the lack of mutants. Perhaps the genes expressed are not present i n the "blank" populations. Most of the seedlings marked f e l l i n Eiche's (1955) "albina" seedling mutant c l a s s , although others, notably those from tree 75, were close to h i s " v i r i d i s " c l a s s . They had bright red hypocotyls and cotyledons with b r i g h t green t i p s and pale yellow bases. They gradually acquired normal colour. In comparison, the mutant from tree 77 had a whitish hypocotyl and cotyledons showing yellow t i p s and pale green bases. Most of the "albino" seedlings were dead within a few weeks of t h e i r emergence, l i k e l y when the food energy i n the seed was consumed. The v i a b l e types gradually became,normal-looking. For instance, that from tree 29 appeared normal by J u l y 12, f i v e weeks a f t e r emergence. These abnormal types d i d not emerge noticeably e a r l y or l a t e . None were marked during the June 3 - 6 inspection, but few seedlings had 97 emerged then. Most abnormal seedlings were found between June 10 and June 27, and most of these were marked on the t h i r d inspection (June 17 -20) when the bulk of a l l germinants emerged. Only two chlorophyll-deficient germinants were found among the progeny of these trees . (one each from trees 37 and 75) when samples from the same trees were incubated on the Jacobsen apparatus i n the laboratory, so the environment i n the nursery apparently had not suppressed t h e i r emergence; they seem equally as able as t h e i r "normal" half s i b l i n g s to muster the reserves i n the seed needed to produce a germinant. This d i f f e r s from the results of Eiche (1955) for Scots pine. Eiche (1955) found that 40% to 55% of the chlorophyll-deficient types found were i n his "xanthoviridis" class. In the present study, none was i n that class, and 11 of 20 (55%) were "albinas", found by Eiche at only 6% i n Scots pine. " V i r i d i s " types formed 40% of the mutants, and the single "xantha" constituted the remainder. These types formed between 11% and 24% of the Scots pine mutants, the l e v e l changing with the test l o c a l i t y . Perhaps environmental conditions here prevented xanthoviridis types from appearing, or my lack of f a m i l i a r i t y with the c l a s s i f i c a t i o n scheme caused m i s c l a s s i f i c a t i o n of xanthoviridis seedlings. Otherwise, i n the populations studied, western hemlock has a greatly d i f f e r e n t genetic structure from that of Scots pine. The frequency of l e t h a l and non-lethal mutant types i s nearly the same at 0.261 and 0.294 percents, respectively. But the value for the l a t t e r i s determined largely by two trees from Mount Seymour, numbers 57 and 75. Tree 75 produced three and tree 57 two mutant seedlings of t h i s type. Cones from the l a t t e r were collected i n the crown base, where s e l f -p o l l i n a t i o n may be greater than higher i n the crown, as i n jack pine (Fowler, 1965a). The high frequency of mutant types from t h i s tree, the highest for 98 any tree at 0.9 percent, supports t h i s postulation, although no values are available for seed collected elsewhere i n the crown. The cones from tree 75 came from high i n the crown and therefore should contain r e l a t i v e l y less selfed seed. In addition, tree 57 produced one l e t h a l albino seedling. Such a combination of l e t h a l and non-lethal phenotypes was not found i n any other case. However, Eiche (1955) found two mutant types i n the progeny of between 21 percent and 34 percent of a l l cone parents, depending on the location of the f i e l d t r i a l . He found up to four mutant types (Albina, Xantha, Xanthoviridis and V i r i d i s ) among the progeny of one seed parent. The frequency of western hemlock parents producing more than one mutant type, 1 of 91 trees, (1.1%) i s much lower than Eiche found. The r e s u l t here for western hemlock suggests that two gene systems are involved. The d i f f e r e n t types might have been produced by s e l f -p o l l i n a t i o n or by cross-pollination between trees similar i n these l o c i , or by a combination of the two. Information on chlorophyll-deficient types produced by s e l f -p o l l i n a t i o n of western hemlock i s presented i n Chapter 4. "Healthy" mutants The frequency and proportion of "healthy" mutants by family appear i n Appendix 6. Western hemlock Dwarf mutants Comparing Tables 3-18, and Appendix 6 , i t i s evident that the "healthy dwarf" types of mutant occur much more frequently than the chlorophyll-deficient types. The number of families involved (49 vs 14), the maximum percentage per family (2.10 vs. 0.90) and the o v e r a l l average percentage (0.24 vs. 0.05) of mutants was higher for the dwarfs than for the chlorophyll-deficient types. The high frequency of dwarf mutants indicated for the progeny of tree 10 i s not r e l i a b l e because i t i s based on only 13 seedlings. The family showing the highest frequency of dwarf types, tree 102, was checked to determine when the abnormal seedlings emerged. Of the 7 mutants, one emerged during the second count, one during the t h i r d count, and the remainder l a t e r . Families 31 and 44 were checked also. Again, some emerged r e l a t i v e l y early, but most were t a l l i e d after the peak of emergence had passed. This suggests they are the products of weak, immature seed, although the seedlings were apparently healthy. Summation by population was conducted as before (Table 3-2 0). No correlation pattern of dwarf mutant proportion with c o l l e c t i n g elevations or l o c a l i t y i s apparent for the Coastal seed sources. For the Nelson c o l l e c t i o n , where the highest elevation produced a high percentage of dwarf types, the low number of surviving seedlings makes the population 100 Table 3-2 0 Frequency and percentage of "healthy" mutant seedlings by parental populations Species No. of No. of M u t a n t t y p e and Elev. par- Seed- Dwarf Bluish Other A l l poo: Area (feet) ents lings No. p, "O No. % No. % % Western hemlock Nelson 2000 3 1170 3 .26 3 .26 .51 3500 2 . 840 1 .12 .12 5200 2 136 1 .73 .73 Haney 100 8 4079 5 .12 .12 500 11 4729 17 .36 .36 1300 11 5596 15 .27 7 .12 1 .01 .41 1800 10 3421 9 .26 13 .38 .64 Seymour 50 9 3711 5 .13 .13 400 8 4323 5 .11 1 .02 .14 1000 3 1410 7 .50 .50 1900 5 1717 1 .06 1 .06 .12 2750 5 2252 12 .53 8 .35 .89 Caycuse 600 6 3057 4 .13 1 .03 .16 Paterson Lake 800 4 1152 3 .26 .26 Mountain hemlock Mt. Seymour 3300 3 1944 106 5.45 5 .45 mean unreliable. The high percentage of dwarfs found for the Seymour 2750-foot c o l l e c t i o n i s due mainly to a single family, number 102, which contributed 7 of the 12 mutant seedlings found. Yet a l l families from that c o l l e c t i o n contributed mutants to the provenance mean, suggesting that the circumstances producing such seedlings i n t h i s environment are f a i r l y widespread i n the area. The time of emergence of these seedlings was checked. No pattern of 101 inordinately early or l a t e emergence, which might have suggested seed damage or immature and weak seed, was apparent except as noted already for tree 102. Thus, although some dwarf seedlings occurring i n such an environ-ment may be caused by immature seed, some of the stunted offspring may be due to genetic combinations alone. "Bluish" mutants The occurrence of "bluish" seedlings i s much less regular. They were found among families from over 1000 feet elevation i n nearly every case. The Coastal co l l e c t i o n s from Haney and Mount Seymour gave the highest pro-portions found, with the lowest Nelson source f a l l i n g below them, but well above the levels f o r the other "marked" populations. The greatest i n f l u -ence on the frequency of these types i s seed parent. Of the 13 bluish seed-l i n g s from the Haney 1800-foot c o l l e c t i o n , 10 came from tree 87. Tree 96 produced f i v e of the 8 b l u i s h seedlings i n the material a r i s i n g at 2750 feet on Mount Seymour. Only two seed parents, trees 44 and 48, produced a l l the mutants from the 1300-foot c o l l e c t i o n s at Haney. The d i s t r i b u t i o n i n the nursery of the mutant seedlings from tree 87 was checked. They were found i n a l l replicates and i n rows from s t r a t i f i e d and u n s t r a t i f i e d seed. Tree 96 gave a s i m i l a r , though less regular, r e s u l t . The apparent independence of these seedlings from environmental i n -fluence and t h e i r concentration i n the offspring of p a r t i c u l a r parents i n -dicates that they are genetically caused. Their concentration i n medium to high-altitude populations from the lower Coast suggests that they r e f l e c t ad-aptation, to such environments, but might represent a detriment to low-al t i t u d e populations. The reverse seems true for the Nelson area, where the lowest c o l l e c t i o n was the only one to produce bluish seedlings. 102 As before, the r e s t r i c t e d number of families and seedlings from the Nelson 3500-foot and 5200-foot collections are inadequate to constitute a s a t i s -factory screening of those populations. More comprehensive sampling and detailed study w i l l be required to determine the frequency, character and possible adaptive role of the tendency to produce bluish offspring. Kung and Wright (1972) postulated that blue-ness i n Rocky Mountain trees was important to high-altitude species i n reducing photo-destruction of auxins by blue l i g h t , which i s stronger at high al t i t u d e s than low down. Perhaps the same applies here to western hemlock. For the coastal c o l l e c t i o n areas, 8 out of 18 families producing bluish seedlings were not frost damaged i n 1972. A l l of these originated at or above 1300 feet at Haney or Seymour. Although other families from the same l o c a l i t i e s showed no f r o s t damage, the tendency to produce b l u i s h offspring may be indicative of above-average fros t resistance, and therefore i s worthy of study under more closely-controlled conditions than attainable i n the seedbeds used. "Odd" mutants One apparently healthy "mutant" was found i n the offspring of seedlings from tree 161 from 600 feet at Caycuse on Vancouver Island. I t was stocky and robust, with sturdy stem and branches, and shorter, thicker, s l i g h t l y curved and more pointed leaves than i t s h a l f - s i b s . The other "odd" mutant found came from tree 51 at 1300 feet i n the Haney Research Forest. I t was more slender and f r a g i l e than i t s half sibs. I t s leaves were sparse, i r r e g u l a r l y disposed, but pointing to the branch t i p , and were softer and more slender than normal. At the end of 3 growing seasons each was noticeably shorter than a t y p i c a l seedling from the same tree and l i k e l y had a poor chance of surviving 103 i n nature. 'These'seedlings are so d i f f e r e n t i n appearance as well as parentage, and so rare i n r e l a t i o n to the others, that no inferences can be drawn about t h e i r frequencies i n t h e i r respective populations and t h e i r adaptive significance. Mountain hemlock The only type of morphological aberration found i n the mountain hemlock families was formation of more than one top, following f a i l u r e of the terminal bud to develop i n the year following i t s formation. Whereas western hemlock seedlings quickly assume t h e i r t y p i c a l habit, with drooping leader and branches disposed i n one plane, mountain hemlock retains an upright axis, with radially-disposed branches. The deformation noted would constitute a strong detriment to the seedlings i n t h e i r need to establish dominance over competing vegetation. With one exception, multiple-topped seedlings were found i n the progeny of a l l cone parents i n each r e p l i c a t e , and i n 17 of the 24 l i n e s sown. The number with multiple tops per row reached 24, averaging 4.4. In view of t h i s consistency i n the nursery and between parents, and the high mean frequency of the "mutant" type compared to the others, i t i s most probably caused by the sub-optimal conditions for mountain hemlock i n the nursery. I t i s a species adapted to growth i n cool, moist conditions. The summers of 1970 and 1971, p a r t i c u l a r l y the former, were f a i r l y dry and hot. Water was applied as apparently necessary, but day-time temperatures often were high, and possibly too high for mountain hemlock. This l i k e l y affected the seedlings' behaviour, perhaps causing malfunction of the terminal bud. A l t e r n a t i v e l y , perhaps those aberrant 104 seedlings had a need for deeper winter c h i l l i n g than was received, pre-venting the terminal buds from moving past true dormancy (Romberger, 1963, pg. 161), or perhaps they responded too quickly to increasing warmth i n the spring and were damaged by a l a t e r f r o s t . Whatever the cause, i t seems most l i k e l y that unfavourable environment i s responsible. Otherwise, mountain hemlock has a very strong genetically-caused tendency to form multiple tops, which would represent a very high genetic load, considering the need to compete successfully i n the severe environment to which i t i s adapted. "Weak" seedlings At the f i n a l inspection i n 1971, those seedlings apparently too weak to survive another year were counted i n each row. The mean proportion of weak and unringed seedlings by c o l l e c t i o n area appear i n Table 3-r21. Table 3-21 Means of weak and unringed western hemlock seedlings, plus  correlation c o e f f i c i e n t , by area of c o l l e c t i o n . Collection Fami- No. of Weak Unringed r S i g n i f i -area l i e s rows /row /row cance Area Elev'n. Mean SD Mean SD Nelson 2000 3 23 0.30 0.88 1.26 1.10 -0. 09 N.S 3500 3 24 0.50 0.78 1.67 1.90 " 0. 61 * * 5200 2 15 0.53 1.30 1.47 2.07 0. 51 * Haney 100 8 64 0.44 0.99 4.00 4.23 0. 05 N.S 500 11 86 0.30 0.74 3.14 3.34 0. 01 N.S 1300 11 88 0.33 0.89 2.32 2.54 0. 03 N.S 1800 10 79 0.37 0.77 2.97 3.48 0. 22 N.S Seymour 50 9 71 0.13 0.41 2.34 2.85 0. 05 N.S 400 8 59 0.29 0.81 1.88 2.15 -0. 07 N.S 1000 3 24 0.25 0.53 2,08 3.50 0. 15 N.S 1900 4 32 0.53 0.72 3.19 3.58 -0. 04 N.S 2750 5 40 0.37 0.67 2.72 2.85 0. 03 N.S Caycuse 600 6 47 0.08 0.28 1.70 2.59 0. 27 N.S Paterson 800 4 31 0.13 0.43 1.90 2.57 0. 47 * Lake 105 I t i s obvious that there i s a very weak association between "weak" and unringed western hemlock seedlings. The only areas for which a s i g n i f i c a n t relationship was found are the high-elevation sources from Nelson, and Paterson Lake. Some of the unringed seedlings were healthy and vigorous from f i r s t emergence, but remained unringed because they f a i l e d to p u l l t h e i r coty-ledons free of the ground 9 weeks after sowing. Others without rings often were extremely weak and eas i l y deformed when t h e i r cotyledons re-mained stuck i n the seed coat. Since t h i s type probably i s the result of immature embryos, a plot of proportion weak over c o l l e c t i o n date was made. No association was found, except that the two Nelson c o l l e c t i o n s mentioned above were collected early i n September, when they appeared to be too immature to pick. The low f i l l e d - s e e d y i e l d and seedling crop per 100 seeds supports t h i s also. However, the Paterson Lake c o l l e c t i o n was made in early December, when the seeds should have been f u l l y r ipe. Support for the independence of proportion weak from c o l l e c t i o n date l i e s i n the absence of a s i g n i f i c a n t "family" term from ANOVA of proportion weak, but significance of i t when the "unringed" data were analysed. Radiographs of seed samples from some collections were examined to compare seed images to the frequency of weak seedlings. Where a s i g n i f i -cant correlation between weak and unringed seedlings had been found, there was agreement with the proportions of incompletely-developed embryos. Plates were not made for a l l l o t s , so a f u l l comparison was not made. However, the independence of weak seedlings at two years i n the nursery from c o l l e c t i o n date, and the apparent correlation with the proportion of immature embryos v i s i b l e on X-ray plates, suggest that differences i n maturity of collections from indi v i d u a l parents, as found by A l l e n (1958a), are responsible for much of the r e s u l t . But since some seedlings consid-ered weak emerged during the second and t h i r d counts, r e f l e c t i n g average or above-average germinative vigor, some weak seedlings may be the r e s u l t of p a r t i c u l a r genetic combinations, rather than seed immaturity alone. Stronger correlations between unringed seedlings and proportion weak might have been found i f separation had been made between those unringed for f a i l u r e to free a cotyledon from the s o i l and those more profoundly weak. But, as recorded here, the proportion of weak seedlings does not shed much l i g h t on the characteristics of the c o l l e c t i o n studied. Mutant summary Five types of aberrant seedlings have been i d e n t i f i e d and analysed. Chlorophyll-deficient, healthy dwarf and "weak'1 types occur apparently randomly i n most or a l l populations from the Lower Mainland area. This suggests they r e f l e c t adaptation to survival i n these environments. The two "odd" seedlings found were far too rare to permit speculation on t h e i r importance. Non-genetic causes, such as immature seed, were indicated as part of the source of "healthy dwarf" and "weak" seedlings. But the fact that both types emerged at a l l times after sowing indicates that p a r t i c u l a r genetic combinations may be responsible for some of the aberrant seedlings found. Cone parent was conspicuously important i n a l l mutant classes.found i n western hemlock. This was most obvious for the bluish and chlorophyll-de-f i c i e n t types, where most parents produced only normal-looking seedlings. 107 No pattern of association between these mutant types was apparent. Thus i t i s possible that d i f f e r e n t genetic systems are involved, and that pro-ducing a high proportion of bluish types i s important i n imparting cold-hardiness, or regulating some other aspect of growth i n colder portions of the range of western hemlock. Since t h i s i s the only "marked" seedling type that seems to have a pattern of d i s t r i b u t i o n i t seems worth recording i n future studies. These seedlings could be tested for frost-hardiness i n r e l a t i o n to t h e i r normal-looking half sibs and to other families from the same l o c a l i t y . Also, more thorough sampling of the Interior populations of western hemlock would increase the chances of detecting similar seedlings and determining whether t h e i r pattern of concentration at higher elevations i n the Lower Mainland i s duplicated i n the I n t e r i o r . Mountain hemlock was represented by only three seed parents, yet they produced f a i r l y large families that contained no chlorophyll-deficient seedlings. In view of Franklin's (1970 ) experience that most species i n the family Pinaceae produce chlorophyll-deficient seedlings, i t i s probable that mountain hemlock w i l l do so also when s u f f i c i e n t populations and individuals are sampled. The one kind of aberrant seedling found i n the mountain hemlock progeny appears due to the sub-optimal conditions for this,' species i n the seedbeds. Nursery survival Survival was high i n both years. The mean value i n 1970 was 93.8% dropping to 90.4% i n 1971. ANOVA of the data for each year showed that only c o l l e c t i o n areas, not trees within areas, was s t a t i s t i c a l l y s i g n i -108 f i c a n t , with p r o b a b i l i t i e s of less than 0.1% and 0.5% i n 1970 and 1971, respectively. The only area with lowered survival was the Nelson c o l l e c -tion from 5200 feet. I t s mean was 82.6% i n 1970, and 77.7% i n 1971. The range of percents for a l l families i n 1971 was from 75%, for tree 10 from 5200 feet at Nelson, to 97.1% for tree 19 from 100 feet at Haney. With so l i t t l e v a r i a t i o n , survival at two years seems an insensi-t i v e measure of differences between provenances or families when reared under these conditions. 109 Bud-opening study The resumption of shoot elongation by the expansion and further d i f f e r e n t i a t i o n of the axis i n a bud i s an important step i n the annual growth cycle. Plants that can make f u l l e s t use of the increasing warmth as spring advances can gain some advantage over those competing for growing space with them, and eventually may be able to suppress them. This would be an advantage i n young seedlings. However, G r i f f i t h (1968) found no association between date of bud burst and amount of terminal or r a d i a l growth of 11 cone-bearing young Douglas-fir during 6 years. In any case, early "flushers" are more exposed to periodic damage to the developing shoot from late spring f r o s t s and may lose a-11 of the poten-t i a l shoot growth through fros t i n exceptional years. Sile n (1964b) emphasised the importance of a close correspondence between the growth periods of a p a r t i c u l a r l o c a l i t y and thafeof an introduced tree population: "...earliness or lateness of vegetative growth i s a l i f e t i m e detriment to a tree planted o f f s i t e . " Consequently, the bud-opening response of plants must be adapted to the cha r a c t e r i s t i c s of the area i n which they occur i f they are to maintain t h e i r positions within l o c a l associations. In order to assess the v a r i a t i o n i n bud burst between and within species and members of a population, observations of bud burst within the progeny of each cone parent were conducted i n the nursery i n 1971. Methods The rows of progeny from s t r a t i f i e d seed i n replicates 2 and 3 were used. Ten seedlings i n each row were selected across the seedbed and marked with a toothpick at t h e i r bases. The spacing between them was as 110 ' uniform as possible. Average, healthy-looking seedlings were chosen. The seedlings were inspected eight times at irregular i n t e r v a l s ; A p r i l 23, May 2, May 6, May 13, May 18, May 22, June 4 and June 15. At each inspection the terminal bud was examined and each seedling was assigned to one of six classes, as follows: Bud Class Bud Appearance 1 "Tight", no apparent swelling 2 Swelling apparent, bud scales not translucent 3 Swelling apparent, scales trans-lucent and leaves v i s i b l e through them. 4 Bud scales separated somewhat 5 Bud scales separated, leaves separating 6 Terminal axis elongating Observations were continued past the stage of bud opening (Class 5) to make certain that the main axis was developing normally, but Class 5 was considered the beginning of bud opening. Data were subjected to analysis of variance of the importance of seed parent and provenance for each date. The mean date of bud opening (bud class 4.55) was estimated for each provenance from a smooth curve drawn through the mean value for each inspection. The number of days required for each provenance to progress from a mean bud stage of 4.1 to 5.0 was read from the same curves. This value represented the rate of bud bursting. I l l Results The study i s weakened by the strong influence on plant behaviour of that portion of replicate - located i n Bed 1 (Figure 3-13). There the plants were noticeably larger and greener the previous autumn, apparently because the s o i l was more f e r t i l e . They were much more advanced i n bud opening at the f i r s t inspections than the same families i n Bed 2 (replicate 3). Comparisons between families for which both s t r a t i f i e d and u n s t r a t i f i e d rows were located i n either Bed 1 or Bed 2 reiterated the strong influence of the d i f f e r e n t conditions i n Bed 1. A " t " test of the mean bud stage of families i n Bed 1 and Bed 2 on A p r i l 25 and May 3 revealed that the plants i n Bed 1 were more advanced than those i n Bed 2. Spot checks of the rows i n Bed 1 and 2 that were not included i n the study (the u n s t r a t i f i e d rows) gave r e s u l t s s i m i l a r to those studied: Bed 1 was fast and Bed 2 was slow. S o i l samples from these contrasting beds were taken i n November, 1971 and analyzed for nitrogen concentration in'the s o i l s laboratory of the Department of S o i l Science of the Faculty of Agriculture, U.B.C. Nitrogen was chosen for determination because the stunted and yellowish seedlings i n Bed 2 suggested a nitrogen deficiency. Ammonium nitrogen was determined c o l o r i m e t r i c a l l y using Nessler's reagent and t o t a l nitrogen by macro - Kjeldahl methods. No connection between nitrogen concentration and seedling response could be seen. The values ranged between 0.0006 and 0.0014% for ammonium and 0.060 and 0.076% for t o t a l nitrogen. Both ammonium and t o t a l nitrogen concentrations were low, but they were s l i g h t l y higher i n a portion of 112 Bed 1 showing poor tree growth and colour than where the trees were fl o u r i s h i n g . And levels of both nitrogen forms were higher i n a portion of Bed 3 that was outside of the experiment and had received no f e r t i l i z e r i n 1971. S o i l composition should not have been responsible for the d i f f e r -ences i n seedling behaviour because a l l of the s o i l had been mixed and loaded into the beds freshly i n the spring of 1970. Perhaps the d i f f e r -ence was due to the extent to which s o i l from the t r i a l of the year before, that had been loaded into the beds below the new s o i l , was within the rooting l e v e l of the germinants. Whatever the cause, i t was not apparent from the analysis done, and i t could not be ignored. Consequently, the following results and discussion are based mainly on r e p l i c a t e 3, which was completely i n Bed 2. The values of mean bud stages by l o c a l i t y , elevation and date of examination for re p l i c a t e 3 are presented i n Appendix 7 . The data from A p r i l 25 and May 3 provide the best chance to detect differences because they present the greatest spread of values. Most analyses of bud-opening stage are based on those dates. Since the bud development classes used here are subjective, rather than objective, the use of ANOVA on them i s not completely proper. In p a r t i c u l a r , any such c l a s s i f i c a t i o n i s weakened by the categorization of a continuous process into steps of equal increment, whereas the stages they represent may not depend on regular increments of environmental influence or plant response. 113 Bud burst stage Mountain hemlock flushed much e a r l i e r than a l l western hemlock i n Bed 2. On A p r i l 25 the mean values for the mountain hemlock families were: tree 105 - 4.6, 106 - 4.7, 109 - 4.3, versus 2.24 for a l l western hemlock families. Comparing them to the western hemlock families collected nearest to them, those from 2750 feet on Mount Seymour, which gave a mean value of 2.2, the slowness of the western hemlock i s apparent. Even those collected from a similar elevation i n the In t e r i o r , 3500 feet at Nelson, were s i g n i f i c a n t l y slower (1.97) than the mountain hemlock, as was the single family (tree 7) from 5200 feet at Nelson. Comparing western hemlock populations, those from the Inter i o r portion of the range did not open sooner than the Coastal c o l l e c t i o n s . To permit a more sensitive comparison, the sources nearest to 2000 feet from Nelson, Haney and Mount Seymour were examined. The trends of development of these provenances appear i n Figure 3-16. (Temperature data obtained from the U.B.C. Plant Science weather station, approximately one-half mile north-west of the nursery and at nearly the same elevation). I t indicates that the Nelson and Haney progeny behaved v i r t u a l l y iden-t i c a l l y , whereas those from Mount Seymour were s l i g h t l y slower. The dates of achieving mean bud opening (a mean stage of 4.55) by the Haney and Seymour populations diff e r e d by only two days (May 8 to May 10). Since Legend: 115 the l a t t e r was the slowest provenance of a l l i n the e a r l i e s t stages, t h i s difference i s not considered indica t i v e of a major difference i n the populations obtained. From the analysis of variance, elevation at Nelson was s i g n i f i c a n t for only the f i r s t inspection. The dates of mean bud opening were May 8.7, 9.0 and 8.0 for the collections from 2000, 3500 and 5200 feet, respectively. I t i s clear that they did not d i f f e r . Elevation was s i g n i f i c a n t as a source of va r i a t i o n i n the Haney material only once: the f i f t h count, taken on May 18th. The two higher collections were s l i g h t l y ahead of the lowest one (5.07 vs. 5.00), but just enough to cause a s i g n i f i c a n t variance r a t i o . Values by parent tree ranged from 4.9 to 5.6, indicating that bud opening was well advanced. Since t h i s inspection i s the f i r s t one for which elevation was s i g n i f i c a n t , i t was considered a chance res u l t . Elevation was s i g n i f i c a n t for the Seymour c o l l e c t i o n from the f i r s t to the s i x t h counts ( A p r i l 25 to May 28), although no consistent pattern with increasing elevation occurred for any inspection. Trees within ele-vation had ceased to be s i g n i f i c a n t by May 28, The sources causing the significance varied, but always included the one from 1900 feet. I t s mean value was 5.48 on May 28 compared to 5.63, 5.77, 5.83, and 5.74 for the collections from 50, 400, 1000 and 2750 feet, respectively. This consistently aberrant population i s represented by fi v e trees whose mean bud stage ranged between 5.40 and 5.60. Those overlapped the ranges of the other c o l l e c t i o n s , so that the population does not appear to behave completely differently from others on the mountainside. This i s supported by the fact that i t started slower than the others, remaining so u n t i l May 13th, when i t was grouped by Duncan's test with the 400-foot and 2750-foot elevations. After May 28 i t did not d i f f e r from any of the others from Mount Seymour. As a check on the possible influence of sample size on the be-haviour of the slow (1900-foot) c o l l e c t i o n , the larger c o l l e c t i o n s from 50 and 400 feet on Mount Seymour and a l l of the Haney co l l e c t i o n s were inspected to see whether or not a sample could be drawn from them to give a mean sim i l a r to the "slow" one under consideration. By selecting the slowest trees from any of the Haney collections i t was possible to obtain 5-tree means no dif f e r e n t from the slow Mount Seymour source, but i t was never possible to do so from the Seymour c o l l e c t i o n s . Sampling the Haney col l e c t i o n s randomly 5 trees at a time gave means di f f e r e n t from the slow Mount Seymour provenance 19 times i n 20 t r i e s , indicating that the apparent difference i n the behaviour of t h i s provenance was r e a l . The positions i n the nursery bed of the rows from the "slow" popu-l a t i o n were checked. Three of the f i v e trees, numbers 58, 60, and 61, were located i n a f a i r l y small portion of the bed. Comparison of t h e i r development with those of trees from the " e a r l i e r " provenances at 1000 and 2750 feet located nearby indicated that location within the bed was not responsible. Within r e p l i c a t e 2, only one of the trees from the slow source was sown i n Bed 1, which weakens the comparisons of i t s behaviour to that of others i n that favourable location. Nevertheless, the tree i n question, number 57, showed the usual response to that bed by being more advanced than the rest at f i r s t count: i t s mean was 4.3, while 117 that of the others was 1.8, ranging from 1.6 to 2.0. Therefore, while t h i s population exhibited a tendency to flu s h later,than others from the same mountain, i t was able to respond s i m i l a r l y to the others i n a more favourable environment. The 1000-foot provenance from Mount Seymour, was faster than the 1900-foot provenance, and often faster than others as w e l l . I t was faster than a l l others at the f i r s t count (April 25), and was grouped with the fastest provenance up to the si x t h count, after which a l t i t u d e and parent tree both ceased to be s i g n i f i c a n t sources of v a r i a t i o n . Unfortunately, t h i s provenance was represented by only three trees, making the value unreliable. As i n the slow provenance from 1900 feet, one tree (53) affected the value of the mean strongly. At the f i r s t count i t had a mean value of 3.7, while trees 52 and 54 had means of 2.2 and 2.5. I t s margin was maintained throughout most of the study and i t was the f i r s t one i n that provenance to reach complete flushing of a l l seedlings. Comparisons of mean bud stage were made between Coastal c o l l e c -tions from sim i l a r elevations. The Seymour source from 50 feet started s l i g h t l y sooner and reached f u l l opening about s i x days e a r l i e r than the Haney material from 100 feet. The Haney 500-foot material started slower than the Seymour 400-foot and Caycuse 600-foot material, but a l l were equal by May 13, p r i o r to complete opening. Bud burst rate The number of days required to progress from an average value of 4.1 to 5.0 are presented f o r each provenance i n Appendix 7 . The mountain hemlock was too advanced at- the: f i r s t inspection to derive t h i s 118 value. No consistent trend between areas or with elevation within area i s evident for the western hemlock populations. For the Haney material, a trend to faster bud opening with an increase i n elevation i s apparent, but not for Nelson and Mount Seymour. The lowest number of days were found for the Nelson c o l l e c t i o n from 3500 feet, the Mount Seymour c o l l e c -t i o n from 2750 feet and the Haney collections from 1800 feet, a l l of which completed opening i n about seven days. The highest number of days to complete bud opening, 14 days, occurred i n the Haney c o l l e c t i o n from 100 feet. This value was affected considerably i n the "slow" May 13 to 18 period. Temperature data for the months of A p r i l and May collected at the U.B.C. f i e l d crop meteorological station were examined. The d a i l y maximum and minimum traces for the period A p r i l 20 - May 30 are presented i n Figure 3-16. The trend of maximum temperature was generally a r i s i n g one u n t i l May 12th, after which a pronounced 10-day cold period began. This corresponds very well to the "slow" May 13 - May 18 period of bud develop-ment. Unfortunately, no examination of the beds was conducted between May 18 and 25, so that the length of this slow period cannot be defined more closely. Another reason for the May 13 - May 18 plateau could be the c l a s s i f -i c a t i o n system. Usually a population mean of 5.0 included trees (families) less advanced and some more advanced than t h i s value. To have a higher mean than 5.0, some seedlings would have had to reach stage 6 - axis elongation. This may require more energy than the progression from stage 4 119 to 5, automatically creating a plateau i n the trend under optimal conditions. Nevertheless, the cold period could accentuate t h i s tendency to form a plateau-between stages 5 and 6. I f the plateau i s largely the r e s u l t of the cold period, the seedlings seem very sensitive to temperature at t h i s stage. Inspection of the curves for each source indicated that the chief difference between the "slow" and the "fast" c o l l e c t i o n s was whether or not they had reached bud stage 5 before the slow period occurred. Those that had, appeared fast; those that lagged only s l i g h t l y behind on May 13, for example Nelson at 2000 feet and Haney at 100 feet, appeared slow. Comparison of the values from si m i l a r elevations revealed incon-sistencies. The low-elevation col l e c t i o n s from Haney and Mount Seymour were di f f e r e n t i n rate with 14 and 9 days respectively. But the former was affected by the slow period, as noted already, whereas the l a t t e r was not. Comparing Haney 500 feet, Mount Seymour 400 feet and Caycuse 600 feet, the l a s t was somewhat faster, at 7.2 vs. 10 days. Two sources near 1000 feet were compared, Paterson (800 feet) and Mount Seymour (1000 feet) gave i d e n t i c a l vaues of 10 days. The values for Nelson at 2000 feet and Seymour at 1900 feet (10 and 11 days) were similar and slower than that from Haney at 1800 feet (7.8 days). But, again, the "slow" period had affected the f i r s t values and not the l a s t one. The one case where elevation consistently achieved significance as a source of var i a t i o n was Mount Seymour. This i s due to an apparent aberration, as a population from a high a l t i t u d e was s i g n i f i c a n t l y faster. 120 The sample of f i v e trees (57, 58, 60-62) included two, trees 57 and 60, whose seed had been collected from the crown base, whereas a l l other cones i n the study had been collected from the upper crowns. Since s e l f -p o l l i n a t i o n i s more l i k e l y to occur i n cones collected low i n a tree's crown (Fowler, 1965a) and since s e l f i n g generally produces weaker i n -dividuals than outcrossing, trees 57 and 60 were compared to the-~rest from the stand. They were neither slower nor faster than the mean for the c o l l e c t i o n at any date. The tree least developed at any date was number 62, from which the cones had been picked high i n the crown, thereby reducing the p o s s i b i l i t y that i t s low vigour was due to s e l f i n g . Of the trees from t h i s stand, tree 62 produced a high number of seedlings surviving the f i r s t winter. The average number of l i v i n g seedlings per replicate for the replicates used are, tree number 57 (31), 58 (81), 60 (46), 61 (40), 62 (72). Thus, the seedlings chosen to represent t h i s tree were not from a r e s t r i c t e d number available. Both replicates from tree 62 were located i n Bed 2 and both were the slowest i n the provenance for the r e p l i c a t e . At no time did the progeny from t h i s tree give a value among the highest for any date; once i t s position as a slow flusher was established i t did not change. The values of a l l families from t h i s provenance located i n Bed 2 were sim i l a r on any date. In view of this consistency of response, tree 62 seems to be genuinely a l a t e flusher, as do the others i n the provenance. Trees as slow as tree 62 were found i n other provenances, for example tree 27 from Haney, so t h i s kind of tree i s not peculiarrrto the Mount Seymour area. 121 Tree 53 was the e a r l i e s t of a l l trees, and since i t occurred i n such a small c o l l e c t i o n i t had a high impact on the provenance mean. Trees nearly as e a r l y as tree 53 were found i n other provenances, f o r example, tree 88 from 1800 feet i n the Haney Research Forest. Since the Haney collectionscontained larger and more uniform sample s i z e s , they were inspected to see whether reducing the s i z e of the samples could produce r e s u l t s as anomalous as those from Mount Seymour, where seedlings from 1000 feet began fl u s h i n g e a r l y and those from 1900 feet were l a t e r f l u s h e r s than the remainder. Although drawing random samples of three f a m i l i e s from each provenance d i d not produce a s i g n i f -icant influence from elevation on bud stage, i t was po s s i b l e to s e l e c t f a m i l i e s d e l i b e r a t e l y whose mean di d d i f f e r from the provenance mean based on the larger sample. Thus, the tendency of provenances d i f f e r i n g by l e s s than 1000 fe e t i n elevation on the same exposure of Mount Seymour to produce s i g n i f i c a n t d i f f e r e n c e s i n bud-opening time i s a r e s u l t of chance i n sampling, and i s not a r e a l d i f f e r e n c e between populations. If a r e a l d i f f e r e n c e i n bud burst threshhold does e x i s t , maintaining i t i n the face of strong gene flow v i a seed and p o l l e n f l i g h t would require steady el i m i n a t i o n of the unsuited types. This could be most surely done i f the locations i n question presented r a d i c a l l y d i f f e r e n t environments, from the surrounding area. A quick check of the vegetation of each c o l l e c t i o n area on Mount Seymour revealed no d i f f e r e n c e i n minor vegetation: a l l areas seemed to be located i n " t y p i c a l " young western hemlock associa-t i o n s . Those from 1000, 1900 and 2750 feet were growing on moderate to steep slopes carrying bouldery g l a c i a l t i l l over bedrock. The two lower 122 colle c t i o n s were on gentler ground and may be from areas of deeper s o i l . Nothing i n the char a c t e r i s t i c s of these s i t e s suggested conditions so demanding that a population of a p a r t i c u l a r type would be favoured. I t i s u n l i k e l y that a true ecotype, as found f o r western white pine by Squillace and Bingham (1957), could have formed here since gene flow up and down the slope by both pollen and seeds i s so l i k e l y . Since parent tree was so consistently s i g n i f i c a n t i n the analysis of the bud stage forthe Haney and Seymour co l l e c t i o n s , the rate of bursting for separate families was determined also. Curves were drawn for each tree i n the 50-foot and 2750-foot col l e c t i o n s from Mount Seymour. The dates of commencing and completing bud opening are presented for i n d i -vidual trees i n each provenance i n Table 3-22. Table 3-22 Days between commencement and completion of bud burst by parent tree i n separate provenances from Mount Seymour Elevation 50 feet Tree Number Days Elevation 2750 feet Tree number Days 65 66 67 68 69 70 71 72 73 LIO:7 11.0 8. 11. 9. 8. .6 ,5 .8 .6 Mean 9.2 6.0 6.1 9.1 96 98 100 101 102 6.0 10.9 10.6 6.8 8.5 8.5 The difference i n the means was tested and found i n s i g n i f i c a n t . The mean values for the provenances derived t h i s way are s l i g h t l y higher than when only the provenance mean value i s plotted, but the r e l a t i v e positions 123 of the means are the same. Each provenance contained families that were quick flushers and others that were slow. Since nearly every family reached a mean stage of 5.0 ( a l l trees f u l l y flushed) on the same date, May 13, the families that were most advanced on the f i r s t inspection were usually the slowest to complete bud opening. Despite the confounding effect of the irregular environmental conditions within the beds, cone parent was consistently s i g n i f i c a n t i n a l l sources but those from Nelson after the t h i r d count. Although bed environment had a strong influence on bud opening, randomization of the seed parents within the bed should have reduced bed influence s u f f i c i e n t l y to permit the va r i a t i o n between families to be analyzed as largely genetically-determined. The strong difference i n flushing response between mountain hemlock and western hemlock i s too great to be explained on the basis of an "ecotypic" difference within a common species as indicated by Langlet (1967) for Scots pine. Judging by t h e i r advanced state of bud developement on A p r i l 25, mountain hemlock possesses a lower temperature threshold for bud opening than does western hemlock. Also, i t was less affected by the conditions i n the nursery beds than western hemlock. The greatest d i f f -erence between the two beds was found i n tree 106 with average stages of 4.4 i n Bed 1 and 4.7 i n Bed 2, respectively. This i s a reversal of the trend found for western hemlock, but the differences are apparently not si g n i f i c a n t . L i t t l e difference was apparent between the parent trees, but t h i s may be due pa r t l y to t h e i r advanced state at the f i r s t inspection. 124 Three parent trees are not s u f f i c i e n t for r e l i a b l e calculations of uniformity or d i s s i m i l a r i t y of behaviour between mountain hemlock fa m i l i e s , but they indicate that mountain hemlock has a sharply d i f f e r e n t bud-bursting behaviour than western hemlock. This study indicates that bud bursting by western hemlock i s under genetic control. Since no observationsof bud opening were taken on the parent trees, i t i s not possible to compare th e i r behaviour with that of t h e i r offspring. I t i s clear that each population contains individuals whose progeny d i f f e r somewhat i n bud-opening behaviour from others i n the stand, as found for other species by S i l e n (1962) and Worrall and Mergen (1967), and that the v a r i a t i o n within a stand can surpass the v a r i a t i o n detectable between provenances separated widely i n elevation or longitude. I t i s clear also that western hemlock i s very sensitive to s i t e d i f f -erences i n i t s bud-opening response, so that a more uniform environment than available here must be used i f bud opening i s to be studied i n the future. 125 Bud set studies The cessation of height growth and the development of a viable terminal bud i s an important phase of the annual growth of woody plants i n temperate climates. Plants that do not respond quickly enough to the s t i m u l i governing terminal bud i n i t i a t i o n w i l l l i k e l y suffer f r o s t damage pe r i o d i c a l l y and lose part of th e i r height growth, putting them i n an i n f e r i o r competitive position as the community closes. The tendency to set buds early and avoid f r o s t s , p a r t i c u l a r l y during the seedling stage, must be balanced by an a b i l i t y to u t i l i z e the growing season as f u l l y as possible and thereby avoid overtopping and exclusion by longer-growing types. Observations of the progeny of widely-differing provenances have i n -dicated that those from cold climates stopped growing e a r l i e r than those from warmer climates. This occurred whether the c o l l e c t i o n s were made poleward or at higher altitudes than the test location. The early experience with population studies of tree species and conclusions of Ortenblad, C i e s l a r , and Engler have been summarized by Irgens - MjzSller (1958) and Langlet (1962, 1967). They found that differences i n be-haviour occur within widely-distributed species and that these are associated with differences i n the climates of the respective areas, producing the "physiological v a r i e t i e s " of Cieslar. Studies have been conducted since to determine the factors c o n t r o l l i n g bud set. Photoperiod has been i d e n t i f i e d as the p r i n c i p a l stimulus to to bud set (Pauley and Perry, 1954, Irgens-MjzSller, 1958, Olson et a l . , 1959) , although i t may interact with temperature (Robak and Magnesen, 1970) . 126 Physical factors (reduced daylength, low temperature, low nitrogen and moisture stress) hastened bud set i n western hemlock (Cheung, 1973). However, genetic factors: l o c a l populations (Hermann and Lavender, 1968), ortet (Pauley and Perry, 1954) and pollen parent (Orr-Ewing, 1966), were apparent also i n bud-setting time of Douglas-fir and western black, Cottonwood (Populus trichocarpa Torr. and Gray). In order to assess the v a r i a b i l i t y i n date of bud set within and between the hemlock populations i n t h i s study, and to relate i t to autumn weather and seedling height, observations of bud set were made i n the nursery during 1970 and 1971-72. Methods The number of seedlings showing a terminal bud was recorded for each row at a number of dates during the f a l l of 1970, the f a l l of 1971, and the winter of 1972. Where the seedling had more than one stem, i t was t a l l i e d as "set" i f one "leader" carried a bud. The number of plants whose terminal bud had opened during the f a l l was recorded for each family. Three re p l i c a t e s , 2, 3 and 4, were observed i n 1970, but only two (2 and 3) could be followed i n 1971 because examination of the t a l l e r trees was much more time-consuming, and because r e p l i c a t e 4 had been damaged by sunscald during 1971, making i t unsuitable for further obser-vation. Replicates 1 and 3 were intended for these observations because each was intact i n a single bed, but too much v a r i a b i l i t y i n survival and plant size existed i n r e p l i c a t e 1 to permit i t s use. 127 Observations began i n 1970 on October 20 and ended December 13. In 1971 they began on October 5 and ended February 3, 1972, because the early, persistent snow cover interrupted the f i n a l count i n l a t e November, necessitating another complete inspection. For rows excluded from the December, 1971, inspection, approximate values of the number of seedlings showing buds were obtained by p l o t t i n g the values for each row and drawing between the November and February values a l i n e similar to that for rows inspected on December 5. The estimates obtained t h i s way are included i n values of Table 3-25, but when included i n analysis of variance and li n e a r regressions, the degrees of freedom were reduced accordingly. The data for each row were converted to the percentage of seedlings with "set" buds at each date. These were converted to arcsin to increase the normality of the variances and the values analyzed f o r family, provenance, region and r e p l i c a t e . The trend between elevations i n the three main areas of c o l l e c t i o n (Nelson, Haney and Seymour) were examined by comparison of regression analysis i n which elevation was the independent variable. Significant results were checked for homogeneity of variance by B a r t l e t t ' s t e s t . Results: 1970 The apparent difference of those portions of replicates 2 and 4 not established i n Bed 2 was tested by comparing the families i n those beds to t h e i r counterparts i n replicate 3. The lower value from Bed 2 material was s i g n i f i c a n t l y d i f f e r e n t for a l l counts. Subsequent analyses were conducted on only the material i n Bed 2, since i t appeared most uniform. The mean values of percentage bud set by area and elevation of seed source based on replicates 2 and 3 i n Bed 2, appear i n Table 3-23. 128 Table 3-23 Mean percentage bud set by area and by elevation of parent tree by inspection date for autumn, 1970 Seed o r i g i n No. of Bud set percentage by inspection date Area Elev. fami-feet l i e s Oct.22 Oct.31 Nov. 14 Nov. 30 Dec.12 Western hemlock Nelson 2000 3 15.5 22.7 40.2 43.7 49.0 3500 3 16.5 21.1 30.0 35.8 50.6 5200 1 42.0 56.4 61.6 71.0 79.3 Haney 100 8 4.0 7.9 ' 18.9 25.8 29.2 500 11 6.2 12.2 27.2 33.2 36.4 1300 11 7.2 14.9 32.3 39.1 45.0 1800 10 7.2 15.9 32.1 38.1 44.5 Seymour 50 9 3.4 8.5 17.8 21.6 32.9 400 8 7.4 12.5 23.1 30.4 33.2 1000 3 2.6 6.3 12.1 18.4 20.7 1900 4 6.3 9.8 14.4 19.5 24.0 2750 5 13.0 18.3 32.6 37.5 40.2 Caycuse 600 6 9.2 14.8 32.4 43.3 50.3 Paterson 800 4 7.5 12.1 17.6 21.4 26.4 Lake Mountain hemlock Seymour 3300 3 100.0 100.0 100.0 100.0 100.0 A l l mountain hemlock seedlings had formed buds at the time of the f i r s t inspection. Further discussion w i l l be based on only the r e s u l t s for western hemlock. One s t r i k i n g feature of the results was the influence of the western sideboards on the number of seedlings that showed buds. This was p a r t i c u l a r l y important i n the early counts, when most or a l l of the seedlings with buds were i n t h i s obvious zone of high bud set. For instance, a l l 6 of the "bud set" seedlings i n the family of tree 4 at the October 22nd inspection were i n t h i s zone of shade from the sideboards. To remove thi s non-random influence, the seedlings within four inches of the western sideboards were excluded from the t a l l y of "bud set" and t o t a l seedlings. Results from the Nelson area progeny were quite consistent with each count, so data from only the October 22nd and December 12th inspections 129 w i l l be discussed. For the other areas, the re s u l t s changed more with inspection date, requiring discussion of the f u l l data. ANOVA of the transformed percentages of data for the October 22nd count indicated that the high-elevation c o l l e c t i o n (5200 feet) from Nelson showed a greater degree of bud set than those below i t . This difference persisted to the l a s t count (December 12), when the untransformed p e r r centages were 49, 51 and 79 percents for the progeny from 2000, 3500 and 5200 feet, respectively. The means from the two lower provenances did not d i f f e r , so that the apparent trend of increasing bud set with increasing a l t i t u d e i s not consistent. Furthermore, the value for the high-altitude provenance i s based only on the progeny of one tree, number 7. Comparing i t s performance to that of other fami l i e s , i t i s s t i l l more advanced at each inspection, but the va r i a t i o n between families within the other Nelson provenances i s obvious. For instance, mean bud set percent for October 22 i n the 2000-foot provenance ranges between 2.9 percent (tree 12) and 39.9 percent (tree 11). These trees stood beside each other i n the forest, indicating that genetical differences might be over-shadowing any environmental influence stemming from the location of the seed parent or progeny row i n the bed. Similar v a r i a b i l i t y was found i n the values from December 12th inspection: tree 11 gave a mean bud set of 64.5 percent and tree 12 only 28.4 percent. For the 3500-foot provenance on October 22, percentage bud set ranged between 7.5 percent (tree 4) and 21.9 percent (tree 2). The results from Haney and Seymour on October 22 showed no influence of elevation. The apparent difference between the collections from 100 feet and 1800 feet was not s i g n i f i c a n t . The same re s u l t was obtained for 130 the material from Mount Seymour, although the percentages were much lower for the 50-foot and 1000-foot material than for the other Seymour co l l e c t i o n s . By the December 12th inspection, both the Nelson and Haney material exhibited a s i g n i f i c a n t response to a l t i t u d e of o r i g i n , but Seymour material f e l l just short of reaching significance. Percentage bud set increased as a l t i t u d e of source increased (Table 3-24). The strongest regression was obtained for the Nelson material, where 40 percent of 2 t o t a l variance was removed by the regression l i n e (r = 0.401). The l i n e for Haney removed only 11.8 percent of the variance, and that for Mount Seymour 7.2 percent. Table 3-24 Linear regressions of arcsin bud set with elevation of seed source for Nelson, Haney and Seymour. Data of October 22 and December 12 , 1970. Seed Elev'n No. of Date Regression feature Source l i m i t s Nur- Slope Inter- Stan. r Sig-Area feet sery .(Arcsin/ cept Err. n i f . rows 1000 f t . ) Nelson 2000- 10 Oct.22 3.18 -0.271 11.920 0.112 N.S. 5200 Dec.12 9.12 -3.329 0.001 0.401 * Haney 100- 55 Oct.22 0.11 1.357 1.691 0.002 N.S. 1800 Dec.12 4.38 6.787 0.080 0.118 * Seymour 50- 43 Oct.12 0.43 0.619 1.742 0.059 N.S. 2750 Dec.12 2.05 5.884 0.074 0.072 N.S. The slopes and intercepts of the regressions from December 12 were compared by covariance analysis, although the variance of the Nelson data was less than those for Haney and Seymour. The trend for the Nelson seedlings was of a steeper increase of bud set with elevation than found for the Haney c o l l e c t i o n , which did not d i f f e r from the Seymour trend. 131 The data from Haney and Mount Seymour were examined to see when si g n i f i c a n t regressions of bud set with elevation f i r s t occurred. Haney material f i r s t showed one from the November 30th inspection. Seymour collections showed s i g n i f i c a n t responses at November 14th and 30th, but not on December 12th. The regressions were compared for s i m i l a r i t y of slope and intercept for each date. The Haney and Seymour regressions did not d i f f e r i n either slope or intercept for any inspection before December 12th, when the Haney l i n e gave a higher intercept. The s i m i l a r i t y of the slopes of the two Coastal c o l l e c t i o n areas indicates that the response to increasing a l t i t u d e of c o l l e c t i o n i n each area i s s i m i l a r . The clim a t i c summaries for those areas (Table 3-4 ) indicate that the e a r l i e s t f r o s t at comparable elevations occurs at nearly the same time. This does not agree with the results of the regression comparison, which indicates that the Haney area.produces higher budset at "sea" l e v e l than Mount Seymour. While i t s mean temperature during early autumn may not d i f f e r much from that of Mount Seymour, the greater exposure to recurrent outbreaks of cold a i r from the In t e r i o r may have required a more f r o s t -hardy population at Haney, necessitating e a r l i e r bud set. In comparison to th e i r counterparts from the Lower Mainland, the collections from Vancouver Island showed consistent behaviour. The Caycuse material (elevation 600 feet) gave values s l i g h t l y higher than those from Haney and Mount Seymour (elevations 500 and 400 feet, respectively). Like-wise, the Paterson Lake material (elevation 800 feet) gave a higher l e v e l of bud set than that from 1000 feet on Mount Seymour. These res u l t s suggest a 132 strong difference i n l o c a l climate not expressed i n the climatic summ-ary (Table 3-4 ), but the Seymour-Paterson comparison i s weakened by the odd behaviour of the Seymour 1000-foot c o l l e c t i o n , which gave sharply lower values of bud set than the c o l l e c t i o n below i t at 400 feet. No association was found between mean height of the container-grown seedlings and the percentage of bud set on December 12. The lack of s i g n i f i c a n t trend of bud set with elevation i n the Mount Seymour progeny i s noteworthy, since bud set, and the concomitant increase i n frost-hardiness, i s such an important phenomenon i n temper-ate climates. The values i n Table 3-23 show that the pattern for December 12 i s erratic::: the c o l l e c t i o n from 400 feet had a higher incidence of v i s i b l e buds than those from 1000 and 1900 feet. A l l calculations were repeated and found accurate. No reason for t h i s anomalous behaviour i s apparent, unless the l e v e l area there i s a f r o s t pocket, thereby selecting a population with s t r i k i n g l y early bud set. Likewise, no reason could be found for the loss of the trend of increased bud set with elevation be-tween November 20 and December 12. Percentage bud set increased for a l l elevations, but apparently e r r a t i c a l l y between c o l l e c t i o n areas. The greater slope of bud set with elevation for December 12 from the Nelson collections indicates that the Coastal and Int e r i o r c o l l e c t i o n s studied possess d i f f e r e n t responses to the conditions i n the nursery. The predicted dates of f i r s t autumn fros t s (Table 3-4 ) at comparable elevations are the same for both Coastal and Interior areas. Yet the s i g n i f i c a n t difference i n slope of bud set with elevation for the Nelson material compared to the Coastal collections indicates that the forces selecting early bud set with increasing elevation near Nelson in t e n s i f y faster there than i n the Lower Mainland area. 133 Results: 1971 As for the data from 1970, v a r i a b i l i t y i n nursery environment appeared to have a s i g n i f i c a n t effect on bud set. This impression was tested by comparing results from the rows of trees i n re p l i c a t e 2, Bed 1, with the progeny of the same trees i n rep l i c a t e 3, a l l of which were i n Bed 2. The percentages were transformed to arcsin values before analysis, and tested for homogeneity of variance by B a r t l e t t ' s t e s t . I n s i g n i f i c a n t non-homogeneity between areas was found for a l l counts i n Bed 2, meaning f u l l analysis of the data i s supportable s t a t i s t i c a l l y . Bed 1 was ahead of Bed 2 for the October 5 count only. However, no difference was found for any count when both rows were i n Bed 2. Therefore, the following r e s u l t s are based on the rows i n Bed 2, making them comparable also to the resu l t s from 1970. Western hemlock Table 3-25 , presents the mean percentage of bud set for replicates 2 and 3, i n Bed 2, by inspection date and area and elevation of seed c o l l e c t i o n . I t i s apparent that bud set increased with each inspection, but p a r t i c u l a r l y between the November and February dates. From ANOVA of the transformed percentages, region, elevation i n 134 region, and trees i n elevations were s i g n i f i c a n t sources of variance. For counts of October 5th and 20th the means for Nelson, Haney and Seymour provenances were a l l s i g n i f i c a n t l y d i f f e r e n t . Since these were not d i r -e c t ly comparable because of the great difference i n mean elevation from the Nelson c o l l e c t i o n s , t h e i r responses were expressed i n l i n e a r regressions which were compared (Table 3 _ 26 ), as i n 1970. Table 3 -25 Mean bud set percentage of western hemlock provenances by inspection date, 1971 and 1972 • Seed o r i g i n No. of Percentage bud set by inspection date famil 1 9 7 1 1972 Area Elev. ies Oct. 5 Oct.20 Nov.12 Dec. 5 Feb. 8 Nelson 2000 3 82.8 87.4 91.4 93.7 96.9 3500 3 84.4 90.6 91.4 96.5 98.8 5200 1 95.8 100.0 100.0 100.0 100.0 Haney 100 8 46.3 62.0 66.3 78.1* 85.7 500 11 51.6 76.6 82.4 89.3* 93.1 1300 11 75.2 82.3 89.8 93.3* 96.7 1800 10 77.4 90.4 92.4 94.6 97.3 Seymour 50 9 35.4 56.7 65.8 71.1* 85.0 400 8 45.1 62.9 70.9 80.6* 89.7 1000 3 40.3 68.8 72.5 84.5* 90.9 1900 4 76.1 87.2 91.6 95.6* 98.0 2750 5 85.1 91.4 94.0 97.1* 98.8 Caycuse 600 7 67.3 79.8 82.6 87.9* 92.1 Pater- 800 4 74.5 87.0 91.0 92.4* 97.8 son Lake * Values at December 5 estimated for some trees from curves of general trend drawn for other trees i n the population. Degrees of freedom for regressions reduced correspondingly. 135 Table 3- 26 Linear regressions of arcsin of percentage bud set with elevation of seed source for Nelson, Haney and Seymour. Data from October 5 and 20, 1971. Western hemlock only. Source Eleva- No. of Date REGRESSION VALUES 2 t i o n rows Oct. Slope Inter- r S i g n i f i -Limits Change/ cept cance ( f e e t ) 1000 f t . (Arc.buds) Nelson Haney Seymour 2000 -5200 100 -1800 50 -2750 10 55 43 5 20 5 20 5 20 7.952 8.768 14.705 14.782 14.251 13.163 36.1353 41.6982 23.9741 39.4908 19.7193 34.6375 0.365 0.402 0.325 0.389 0.697 0.558 N.S. * ** ** ** ** The Nelson material did not show a s i g n i f i c a n t trend of bud set with a l t i t u d e for the f i r s t count, whereas those from Haney and Seymour did. These l a t t e r regressions were compared f o r s i m i l a r i t y of slope and intercept by covariance analysis. The only s i g n i f i c a n t difference found was between intercepts for the October 20 inspection when Seymour gave a lower 2 value. The Seymour results gave a better f i t than those from Haney: r of 0.56 vs. 0.39. The material from the Nelson sources displayed no response with elevation of c o l l e c t i o n after October 20, but the co l l e c t i o n s from Haney and Seymour gave s i g n i f i c a n t regression values for a l l inspections. Comparisons of these regressions showed that the trend of increased bud set with increased elevation of coll e c t i o n s (slope) did not d i f f e r between areas for any inspection. However, as before, the intercepts dif f e r e d for the inspections of November 12 and December 5, the Haney material giving a higher value i n each case. The regression for Nelson progeny on October 136 20 gave a s i g n i f i c a n t l y lower slope than those from Haney and Seymour. The col l e c t i o n s from Caycuse and Paterson Lake on Vancouver Island were compared to those from s i m i l a r elevations i n the Lower Main-land area. Caycuse (600 f t . ) was ahead of both i t s counterparts (Haney 500 f t . and Seymour 400 f t . ) on October 5th, and s t i l l ahead of the Seymour c o l l e c t i o n on December 5th (Table 3-25 ). I t was equal to the Haney material on October 20th. The Paterson Lake progeny from 800 feet was wel l ahead of i t s Lower Mainland counterpart (Mount Seymour, 1000 ft . ) at every inspection. This Seymour material was behind i t s neighbouring collections from 400 and 1900 feet on October 5th, but had f a l l e n into l i n e somewhat by the October 20th inspection. The f i r s t inspection i n 1971 apparently was too lat e to detect the f i r s t bud setting i n the Nelson c o l l e c t i o n , and so determine whether e a r l i e r bud setting occurs i n t h i s area. A better estimate w i l l be available from the plantations established i n the U.B.C. Research Forest at Haney, where more families are included f o r t h i s and many of the other provenances. However, from the f i r s t count (October 5th) to the l a s t i n 1971 (December 5th) bud setting degree i n the Nelson progeny from 2000 feet did not d i f f e r from that of i t s comparable collections from the Lower Mainland; Haney (1800) feet and Seymour (1900) feet. The indication of s i m i l a r i t y of bud-setting response i s strong, and indicates a s i m i l a r i t y i n climate for the late summer and f a l l i n these areas. This d i f f e r s from the trend of higher bud set per unit of elevation found for the Nelson material i n 1970 although e a r l i e r inspection might have shown the steeper trend with elevation found for Nelson progeny i n 1970, 137 When compared to the re s u l t s for the same rows i n 1970, strong differences are apparent. Higher values of bud set are found for a l l c o l l e c t i o n s at each date i n 1971. No s i g n i f i c a n t trend of increased bud set with elevation for a comparable date i n October was found i n any case i n 1970, whereas they exist for the Lower Mainland material i n 1971. This may r e f l e c t the increased ease of seeing buds on the larger seedlings, rather than a difference i n response pattern i n the two years. To check on the possible importance to bud set of differences i n weather, temperature data for the period September 1 to December 31 were obtained for both years from the U.B.C. weather station. Grass minimum temperature, rather than screen minimum, was studied because i t should r e f l e c t more closely the temperature at seedling l e v e l . Despite the sharp flu c t u a t i o n i n temperatures, there was a f a i r l y strong general s i m i l a r i t y between the years (Figure 3-17 ). 1970 was about one degree Fahrenheit colder when judged by mean monthly minimum but at times the values for either year were lower than for the other. 1970 was colder i n early September and during late September - early October and i t d e f i n i t e l y was colder i n the l a t t e r part of November and during early December. Both years showed a warming trend i n l a t e October to mid-November. Although 1970 was only s l i g h t l y colder than 1971, i t produced lower bud set i n the same bed. This may be a res u l t of what Magnesen (1969) found i n his lower-temperature environment: bud "set" appeared delayed because the reduced growth of the seedlings meant smaller buds and l a t e r separation of the needles to make the bud v i s i b l e . This i s the l i k e l y reason for the s i g n i f i c a n t difference between Beds 1 and 2 i n 1970. 60 S e p t e m b e r O c t o b e r N o v e m b e r ' D e c e m b e r F i g u r e 3-17 T r a c e s o f g r a s s minimum temperature °F f c r p e r i o d Sept. 1 - Dec. 31, 1970 and 1971. co 139 Another difference between the 1970 and 1971 results was found for the Haney and Seymour areas. Each produced a s i g n i f i c a n t regression of increased bud set with elevation i n 1971, but only Haney did i n 1970 (December 12). The slopes of these regressions for October 20, 1971 d i d not d i f f e r , indicating that the rate of response with a l t i t u d e was the same, despite the proximity of the Seymour area to the moderating influence of the sea. However, the intercept value was higher for the Haney re-gression for the inspection dates of October 20 to December 5. This indicates that the material from the Haney area begins bud set e a r l i e r than that from Seymour, and l i k e l y that the mean date of f i r s t f r o s t i s e a r l i e r at Haney than Seymour, despite the near uniformity i n Table 3-4. A further difference from 1970 i s the lower trend of bud set with eleva-t i o n from Nelson as compared to the Haney and Seymour sources. This i s the reverse of the resu l t from the 1970 data of December 12 and decreases the l i k e l i h o o d that the forces affecting the timing of bud set per unit of elevation change more rapidly i n the Nelson area than i n the Lower Mainland. E a r l i e r inspection i n 1971 might have produced a resu l t similar to that i n 1970. Family (seed parent) effects were important each year, indicating that intra-provenance v a r i a t i o n i n t h i s feature i s considerable. At each inspection, the tree showing highest bud set was number 7, from 5200 feet at Nelson. That showing least bud set was number 68, from 50 feet on Mount Seymour. Within each provenance, certain families were fast and others were slow to form buds. For instance, within the Seymour 50-feet provenance just mentioned, trees 69 and 73 had attained bud set percentages of 71,0 and 74,8 by October 20th. They equalled the mean bud set for 140 the 1000-foot provenance on Mount Seymour for the same date. S i m i l a r l y , the 1800-feet provenance from Haney contained trees (numbers 86 and 89) not di f f e r e n t from the mean for the 500-foot provenance. Despite t h i s strong within-provenance v a r i a t i o n , i t i s p l a i n that there i s a d e f i n i t e trend of increased bud set with increasing elevation of seed source within a very r e s t r i c t e d sampling area. This i s a much stronger trend than that for height or height -f diameter discussed e a r l i e r , indicating that bud setting i s a.much more closely-regulated phenomenon than i s height growth. The general relationship between mean height of the plug-grown seedlings and bud set i n October 20, 1971, i s presented i n Figure 3-18. Although great scatter i s evident, a highly s i g n i f i c a n t l i n e a r r e l a t i o n -ship of decreased height with increasing bud set was found. However, 2 i t s c o e f f i c i e n t of determination (r ) was only 0.102, meaning that bud set percentage explains only 10.2% of the v a r i a t i o n i n height between families. This i s apparent i n the difference i n height between the values for trees 25 and 37. Their bud set levels are about the same (37% and 39%), but t h e i r mean heights are 19.66 and 11,62 cm., respectively. S i m i l a r l y , trees 77 and 10 gave the shortest western hemlock families i n the study at 10.20 and 10.95 cm., but t h e i r bud sets are 31.7% and 90.0%, respectively. Regressions r e l a t i n g these two characters were calculated for each region (Table 3- 27 ). 141 20 19 18 17 0 16 +J -rH 15 a 14 13 12 11 10 25 65 O o o + o Legend A Nelson 0 Haney • Seymour + Caycuse • Paterson 25 = Family number D 165 + O p A o &> Q D 037 O 48 10 77 10 20 30 40 50 60 70 80 90 A r c s i n ( P e r c e n t b u d s e t ) t o O c t o b e r 20, 1971 Figure 3-18 Association between mean height of plug-grown seedlings and percentage bud set in nursery to October 20, 1971. 142 Table 3- 27 Seed Source Linear regressions of the form: Plug seedling mean height (cm.) = a + b (Bud set Oct. 20, 1971) by area of seed o r i g i n , western hemlock families only. Intercept (cm.) Stan error 2 r Signif 19.857 1.578 0.450 N.S. 18.763 1.862 0.218 * * 15.786 1.797 0.082 N.S. 12.618 1.260 0.128 N.S. 16.629 1.816 0.102 * * Nelson 7 Haney 40 Seymour 30 Vancouver Island 10 A l l pooled 87 -0.0884 -0.0773 -0.0313 0.432 -0.0404 Only Haney material gave a s i g n i f i c a n t relationship. Nelson pro-geny gave a strong slope and a regression that approached significance, but the small sample lim i t e d the v a r i a b i l i t y allowed. Neither Seymour nor Vancouver Island material suggested an association between family mean height and bud set. Thus, bud set and height growth appear only weakly correlated using heights of the plug-grown material. Perhaps measurements on the same nursery seedlings observed for bud set would have revealed stronger relationships. The agreement between bud set and height growth within a row was obvious for the early counts i n 1971, Seedlings showing buds were the shorter ones i n every case. Often they were next to a t a l l e r seedling apparently s t i l l i n active growth. For l a t e r counts the agreement was not so strong, since many t a l l seedlings showed buds but some shorter ones did not. When the t i p leaves were separated to check for a bud on those seedlings appearing budless, nearly every seedling had one. This i n -dicates that the c r i t e r i o n for bud set used here i s not s u f f i c i e n t l y 143 sensitive. I t was used to allow comparison between 1970 and 1971 results and to avoid possible interference with bud set i f the terminal axis was affected by separating the leaves a number of times. Provided such an effect i s not produced, bud set could be followed better i f a standard number of seedlings per row were selected and scored p e r i o d i c a l l y for bud stage. This would reduce the possible intra-row influence of differences i n n u t r i t i o n or drainage and change the nature of the measure-ments from binomial (presence-absence) to a more normal (degree of development) d i s t r i b u t i o n , at least for that portion of the response where var i a t i o n i n stages i s greatest. Although no conclusion about the factors inducing bud formation i n t h i s species can be drawn from t h i s phase of the nursery study, the consistency with which those western hemlock seedlings closest to the western sideboards formed buds early and remained short may be an i n d i -cation that temperature interacts with photoperiod to induce bud set, as found by Robak and Magnesen (1970), and suggested by Olson et a l . (1959) for eastern hemlock. The apparent response here i s very marked. If i t i s determined p a r t l y by temperature, the s e n s i t i v i t y of western hemlock to temperature seems very high. Another possibl&explanation i s s o i l moisture differences. Water was sprinkled onto the beds from a l i n e of pipe to the west side, creating a s l i g h t rainshadow near the western sideboard. This may be to blame for the high bud set i n that zone, as found by Cheung (1973), rather than low temperature or reduced l i g h t strength. I t i s notpossible to decide which set of results T- from 1970 or 1971 -represents the bud-setting response of these populations most accurately. 144 Those of 1970 show low bud set and l i t t l e influence of elevation, but a strong influence of parent tree on seedling behaviour, whereas the 1971 results show a d e f i n i t e influence of elevation of seed source, as well as an important influence of seed parent. These results support those of the B r i t i s h provenance study that bud set increases with a l t i t u d e of seed source, and that the Interior source set buds early, but d i f f e r i n the weak r e l a t i o n between bud set and height. No doubt the use of material grown i n d i f f e r e n t l o c a l i t i e s and conditions, and the use of famil i e s , rather than provenances, are i n part responsible for these inconsistencies. As emphasized by Dietrichson (1964), Robak and Magnesen (1970), and Hagner (1970a) annual rhythm i s a very sensitive expression of con-dit i o n s of the o r i g i n of a seed source, since i t i s by the degree of control over the thresholds regulating bud opening and bud setting that a population can maintain i t s e l f i n an area i n the face of competition from other organisms. The 1971 r e s u l t s of t h i s study indicate that there are d e f i n i t e differences i n the growing conditions with increasing a l t i t u d e of western hemlock seed c o l l e c t i o n and that the l o c a l population of ccne«bearing trees i s genetically adapted to the climate i n each area, although there apparently i s strong intra-population v a r i a b i l i t y i n the preciseness of t h i s adaption, or of i t s transmission to the next generation. Mountain hemlock Casual examination of the progeny i n mid-June, 1971, revealed that bed set had occurred already i n many of the mountain hemlock. Since t h i s was noted on June 15th, before the longest day of the year, bud set here appeared to function contrary to the general theory of photoperiodic control of bud set. Consequently, a more thorough examination was con-145 ducted. The number of seedlings possessing terminal buds and the number whose i n i t i a l terminal bud had flushed again was recorded for a l l l i n e s of progeny i n each replicate on the following dates: June 15, June 27, July 9, and July 20. No further examinations were made u n t i l the autumn because nearly a l l seedlings, excluding a few with tops damaged by insects, had set buds by the l a s t date. Table 3-28 presents the data on percentage bud set and secondary flushing for each tree by date of inspection. The sums of the values for buds set and buds re-opened do not equal 100 because the l a t t e r i s i n -cluded i n the former. The minimum number of trees i n a row was 67; the maximum was 97, so the percentages were calculated on substantial numbers and can be compared with reasonable confidence. Table 3-28 Percentage of mountain hemlock seedlings with set and re-opened buds by inspection date and parent tree. Parent June 15 June 25 July 9 July 20 Oct, 4-6 Tree Percentage of buds set re-op set re-op set re-op set re-op set re-op 105 90.8 10.3 97.8 13.3 98.6 9.6 100 2.2 100.0 0 106 94.1 8.2 98.8 5.2 98.6 4.4 100 1.1 100.0 0 109 93.8 8.7 97.0 7.8 98.2 4.3 100 1.0 100.0 0 Very l i t t l e difference between the progeny of the separate parent trees was apparent. Tree 105 showed a lower bud set than the others on June 15, but equalled them by July 9. I t had the highest percentage of reopened buds at each inspection and reached a peak value l a t e r than the 146 other families. By October 4 to 6, when the f i r s t f u l l inspection of the western hemlock families was made, no re-opened buds remained. Here, no influence of nursery bed was apparent, nor did analysis of variance of the percentages transformed into arcsin indicate one. The behaviour of mountain hemlock i n setting buds so early i s greatly d i f f e r e n t from that of western hemlock. The progenies of high-a l t i t u d e western hemlock co l l e c t i o n s were examined at the same time. No terminal buds were found. Many of the mountain hemlock buds observed on June 15th apparently were temporary ones that opened after a br i e f pause to permit more growth before a f i n a l bud was set. Thus, height growth i n mountain hemlock i s by a series of flushes of temporary buds, rather than of uninterrupted elongation of the axis. The apparent independence of the appearance of the f i r s t buds from photoperiodic control i s noteworthy. I t indicates that control of bud setting i n these two hemlock species i s by diff e r e n t mechanisms. Kozlowski (1971a) indicates that some woody plants can enter dormancy under long days. Whether or not these seedlings had reached a dormant state, t h e i r height growth had ceased i n a l l but a few cases before the longest day'of the year. Perhaps unfavourable temperatures overrode the i n f l u -ence of photoperiod on bud set as found by Olson et_ al.(1959) for eastern hemlock. Since mountain hemlock occupies cool temperate areas, i t should be adapted to continuing growth under cool conditions. May and June of 1971 were generally cool and wet and should have offered i d e a l conditions for the development of mountain hemlock seedlings, provided the pattern and degree of day-night temperature differences i n r e l a t i o n to that i n the 147 native sites was not too different. Much higher temperatures occurred during the summer of 1970, but no comparably early bud set was noted. If the same earliness of bud set i s found in successive years, i t w i l l offer a dramatic way to screen populations of putative western hemlock-mountain hemlock hybrids. Under similar conditions, they should show high bud set in mid August to early September before many western hemlock seedlings, even those from high-elevation sources, have vi s i b l e buds. 148 1972 Frost damage Methods Frost damage was noted throughout the nursery beds i n February of 1972. A l l rows were inspected and the number of damaged seedlings re-corded for each. Seedlings showing bright red-brown foliage at or near the t i p were entered into one of the following damage claeses: Class Description Appearance 1. Undamaged No damage 2. Light Tip leaves damaged along the margins or not beyond half the leaf length, terminal bud considered undamaged. Heavier damage elsewhere on the plant possible. 3. Medium A few t i p leaves damaged to near base; terminal bud l i k e l y a l i v e . 4. Heavy Many t i p leaves damaged to bases; terminal bud l i k e l y dead. The frequency i n each damage class i n each row was m u l t i p l i e d by a weighting factor. These were: 0 for class 1, 1.0 for class 2, 2.0 for class 3, and 4.0 for class 4. These products were summed and the sum divided by the number of trees i n that row. This value, i n turn, was transformed by arcsin to normalize the variance before analysis of the influence of r e p l i c a t e , area and elevation of o r i g i n , and parent tree was conducted. The homogeneity of the variances of t h i s transformed value was tested by B a r t l e t t ' s method, The data were analysed to detect patterns associated with provenance, seedling height, "plug" f r o s t damage, and bud set. 149 Results: Influence of replicates Replicate 1 was damaged s i g n i f i c a n t l y more heavily than the others, so i t was eliminated from the main analysis. Within re p l i c a t e 1 no s i g n i f i c a n t influence from area or trees i n stands could be detected i n the o v e r a l l damage. Apparently i t was due to a strong environmental influence, tempered only by seed source elevation. The mean values of the untransformed, weighted damaged found: i n rep.li-v • cates 2, 3 and 4 are presented by area and elevation of seed source i n Appendix 8 , and Figure 3-19. Mountain hemlock was undamaged. ANOVA of the western hemlock data indicated that area, elevation and parent tree a l l affected the damage rating. Influence of provenance The o v e r a l l damage to Seymour progeny was higher than from the Nelson and Haney sources, which did not d i f f e r . Inspection of Figure 3-19 shows that the three lowest Seymour sources were damaged much more heavily than a l l others. Damage to the Haney progeny decreased regularly with increasing a l -titude above 500 feet, reaching as low a l e v e l as that of the Nelson material from 2000 and 3500 feet. The Nelson c o l l e c t i o n from 5200 feet showed an i n -crease i n damage over i t s two Nelson counterparts from lower elevations. The col l e c t i o n s from Vancouver Island, (Caycuse and Paterson Lake) appeared to be less damaged than material from comparable elevations on the Lower Mainland, This was checked by ANOVA and found s t a t i s t i c a l l y true i n each case. However, the Caycuse and Paterson Lake co l l e c t i o n s did not d i f f e r i n damage rat i n g . The v a r i a b i l i t y i n damage rating within provenances i s expressed as standard error of the mean i n Appendix 8 . I t shows a general decrease i n v a r i a b i l i t y with increase i n elevation for the Haney and Seymour material. Legend 2000 3000 Elevation of source (Nelson) ' 5000 Figure 3-19. Patterns of nursery fro s t damage with seed source and elevation. Replicates 2,3, 4 only. ^ j_j Ul o 151 Data for Haney were most variable below 1300 feet, whereas the Seymour results varied most at 1000 feet. This may be due p a r t l y to the low number of families (3) available. For the same reason the standard error of the Nelson 5200-foot c o l l e c t i o n l i k e l y was high since i t was based on only 2 families. Influence of elevation of source Despite the fact that variances were s i g n i f i c a n t l y d i f f e r e n t with elevation i n each area, lin e a r regressions of damage rating on elevation were calculated for the material from the Mount Seymour, Haney and Nelson co l l e c t i o n s , to see whether the apparent trends of damage with elevation were d i f f e r e n t . The values for these l i n e s appear i n Table 3-- 29 -Table 3- 29 Linear regressions of 1972 f r o s t damage rating with eleva-t i o n for progeny of Seymour, Haney and Nelson areas. Replicates 2, 3 and 4 only. Damage Basis: per seed- 1000 f t . Stand-li n g s elev'n. Inter- ard 0 Region (Rows) r i s e cept error r Prob. Nelson 1,655 0.006 -0.0137 0.0344 0.044 N.S. (46) Haney 13,637 -0.023 0.0438 0.0651 0.064 * * (240) Seymour 10,378 -0.027 0.0807 0.0947 0.080 * * (171) The trend from the Nelson material was of an i n s i g n i f i c a n t increase of damage with increasing elevation, while i t was the opposite from the other material. This anomalous r e s u l t from Nelson i s based on only one 152 heavily-damaged seedling i n a row of 17 progeny of tree 7. The other Nelson seedlings were damaged only s l i g h t l y ("light" damage). They occurred i n the progeny of trees 12 and 14, from 2000 f t . Therefore, t h i s trend i s very weakly based and should not be considered i n d i c a t i v e of the r e l a t i v e f r o s t resistances of the populations u n t i l longer-term observa-tions are obtained. The Haney and Seymour regressions were compared for s i m i l a r i t y by covariance analysis. Their intercepts, but not thei r slopes, differed s i g -n i f i c a n t l y ; that from Seymour was high (Table 3- 29 ). Although the c o l l e c -tions from these areas spanned similar elevations, t h e i r f r o s t suscepti-b i l i t y appeared d i f f e r e n t . This was examined further by comparing the ratings for col l e c t i o n s from similar elevations. For a l l elevations com-pared (100 vs. 50 feet; 500 vs. 400 feet, and 1800 vs. 1900 feet i n the order Haney, Seymour) Haney material sustained less damage than Seymour progeny. Thus, the indication of greater f r o s t s u s c e p t i b i l i t y of Seymour material indicated by the regression analysis was supported consistently by these comparisons. Family influence The importance of parent tree on the provenances' damage rating was tested i n a l l ANOVAS. I t was usually highly s i g n i f i c a n t . I t was appraised further by inspection of the weighted family means from the analysis of variance for replicates 2 to 4 pooled. They showed great v a r i -a b i l i t y . The heaviest damage occurred i n the progeny of tree 77, which had a rating of 0.309. I t i s located at 400 feet on Mount Seymour. Every other family i n t h i s c o l l e c t i o n suffered some damage, the ratings ranging from 0.027 to 0.080. Most of the other co l l e c t i o n s up to 600 feet showed 153 damage i n every family (Appendix 8). The exceptions were the ones from 50 feet on Mount Seymour, where tree 67 was damage-free, and from 600 feet at Caycuse, where tree 157 escaped damage. As the elevation of source rose, damage incidence and severity dropped, as indicated already by the significance of the damage-elevation regressions. However, as found for the Seymour c o l l e c t i o n from 1000 feet, t h i s trend was not consistent. Strong differences occurred between areas also, as indicated i n Figure 3-19. At 1900 feet on Mount Seymour, 4 out of 5 families were damaged, whereas at a comparable elevation i n Haney (1800 f e e t ) , the material showed damage i n only 2 out of 10 families. The severity of damage was less i n the Haney material also, as shown above, only " l i g h t " (Class 2) damage occurred there, whereas "medium" (Class 3) and "heavy" (Class 4) damage was found i n the Seymour material. This i s reflected i n the damage ratings, which were 0.0011 for Haney and 0.0179, 18 times higher, for the Seymour progeny from 1900 feet. The highest c o l l e c t i o n from Mount Seymour, at 2750 feet, had damage i n 2 of 5 families. The severest damage found there was that one seedling of a t o t a l of 5 damaged, was i n the "medium" class. Relation to height growth The relationship between mean height of the container-grown material and the f r o s t damage of the same families i n the nursery was investigated. From Figure.3-20 , the pattern seems confused. Both t a l l f amilies (cone parent 65) and shorter ones (cone parent 73) from the same stand were damaged about equally. The most severely-damaged family, 77, was the shortest one i n the container-grown material. The t a l l e s t family, number 25 from 100 154 E O tn c •H •a cu cu U) C o Cn l Cn . 3 rH ft «H O Cn •H rC C rt Qi X 20 19 18 17 16 15 14 13 12 11 10 O 25 65 O 19 o o A A o 6 o o o 6<P Legend: A Nelson O Haney Seymour + Caycuse • Paterson r = 0.013 d.f. 87 N.S. 73 77 0.1 0.2 0.3 F r o s t d a m a g e r a t i n g n u r s e r y Figure 3-20. Relationship between mean height of plug-grown seedlings and nursery f r o s t damage for western hemlock families. 155 feet at Haney, was damaged much more l i g h t l y . Undamaged families ranged i n height from 11.35 cm. (tree 48 from 1300 feet at Haney) to 16.90 cm. (tree 157 from 600 feet at Caycuse). The relationship between mean height and f r o s t damage was tested by regression analysis for a l l data pooled and for i n d i v i d u a l areas (Appendix 9 ). No area produced a s i g n i f i c a n t relationship, nor was an o v e r a l l trend found. In comparison to these results from replicates 2, 3 and 4, those from r e p l i c a t e 1 (Appendix 9 ), show the following differences: a highly s i g n i f i c a n t relationship existed between seedling mean height and weighted f r o s t damage; only elevation (not area or cone parent i n the stand) was s t a t i s t i c a l l y s i g n i f i c a n t ; a few extra families were damaged (number 11, 67 and 165), but many families formerly damaged escaped here (numbers 30, 31, 47, 57, 60,61, 85, 91, 101, 153). In Figure 3-21, the general position of points above the l i n e of perfect agreement between replicates 2 to 4 combined and 1 i s apparent. This demonstrates the generally heavier damage i n the l a t t e r r e p l i c a t e . The most heavily-damaged family i n replicate 1 was number 35, rather than 77. Their damage ratings were, respectively, 0.089 and 0.309 for replicates 2 to 4, and 0.543 and 0.390 for replicate 1. The damage to family 35 occurred completely i n one of the two rows available. Fifteen of the nineteen seedlings damaged were i n the "heavy" class; two others were considered "medium". This was the highest frequency and severity of damage found i n any row i n the nursery. Tree 25 also suffered much damage i n replicate 1, the rating r i s i n g from 0.051 to 0.485, mostly because of a sharp increase i n the "heavy" damage class (12 out of the 23 damaged i n r e p l i c a t e 1 as opposed to 2 of 14 seedlings i n the others). Figure 3-21 Comparison of weighted f r o s t damage ra t i n g of Replicate 1 vs. Replicates 2-4 Ln 157 Those families not damaged i n rep l i c a t e 1 were often those damaged only l i g h t l y i n the other r e p l i c a t e s , although 22 families showed no damage i n any replicate (Table 3-30 ). The Caycuse c o l l e c t i o n from 600 feet was the lowest one to include an undamaged family. The Nelson material from 2000 feet appeared less frost-hardy than that from Haney at 1800 feet. Table 3-30 Origin and family numbers of progeny showing no fr o s t damage i n the nursery, February 1972. Area Elevation Number of families Family numbers (feet) Total Undamaged Nelson 3500 3 3 2, 4, 5, 5200 2 1 10 Haney 1300 11 3 45, 46, 48, 1800 10 8 86, 87, 88, 89, 90, 92, 93, 94 Seymour 1900 5 1 62 2750 5 3 96, 98, 102 Caycuse 600 6 1 157 Paterson 800 4 2 161, 162 Families damaged for the f i r s t time i n replicate 1 include the one showing the highest damage proportion of a l l , tree 67, where 7 of 31 seedlings (22.6%) were frosted. The small sample available made the impact of those damaged very strong, producing a rating of 0.375. In other r e p l i c a t e s , t h i s family was represented by a t o t a l of 97 trees i n 6 rows, meaning that the lack of damage there was reasonably based. The regression of weighted fros t damage i n replicate 1 on the height of plug-grown seedlings d i f f e r e d from that of replicates 2 to 4 i n having a s i g n i f i c a n t relationship, with a stronger slope, and having a negative intercept, meaning that damage should occur, whether or not a "basal" seedling 158 height was reached. Relation to bud set The general relationship between f r o s t damage and bud set i n western hemlock families on October 20, 1971, i s presented i n Figure 3-22. Despite a strong scatter of damage rating for families showing low bud set, a highly s i g n i f i c a n t l i n e a r relationship (probability less than 1%) ex i s t s . Using bud set on October 5, 1971, an equally strong lin e a r relationship was found. Neither slope (-0.00152 damage units per percent bud set) nor intercept (0.0928 damage units) d i f f e r e d from those presented for the October 20 data. Further discussion w i l l be limited to results from October 20. I t i s clear that those families showing the highest percentage of budset on October 20 tended to suffer less f r o s t damage. But i t i s clear also that there are some s i g n i f i c a n t l y aberrant families. For example, trees 65 and 68, from the same c o l l e c t i o n at 50 feet on Mount Seymour, have approximately the same degree of bud set, but the former suffered much higher f r o s t damage. And tree 73, also from the same source, was about as frost sensitive as 65, although i t had a considerably higher bud set. As mentioned previously, tree 67, also collected there, showed no damage despite showing very few buds. Tree 7, from 5200 feet at Nelson, had a high damage rating from only 1 frosted seedling because the number of established seedlings was so low. Regressions of f r o s t damage to western hemlock i n replicates 2 - 4 , and 1, on percent bud set observed i n October 20, 1971 were calculated for each c o l l e c t i o n area (Table 3-31). 159 Legend A Nelson O Haney » Seymour 160 Table 3-31 Regressions of the form: Frost damage = a + b (Bud set Oct. 20, 1971) for areas of western hemlock seed sources: A - Reps 2 - 4 2 Seed Sample Slope: dam- Intercept: Stan. r Proba-source size age units / damage error b i l i t y % bud set units x 1000 Nelson Haney Seymour Vancouver Island 7 40 30 10 0.000649 -0.001212 -0.001870 -0.000108 -0.03741 0.08943 0:14163 0.01516 0.33 0.436 0.32 0.271 0.68 0.213 0.40 0.009 N.S. * * N.S. 0.08791 1.25 0.114 N.S. 0.36178 1.49 0.221 * * 0.35804 1.15 0.402 * * 0.15012 1.42 0.194 N.S. The collections from Haney and Mount Seymour showed s i g n i f i c a n t reduction i n f r o s t damage with an increase i n v i s i b l e buds i n both cases, while those from Nelson and Vancouver Island did not. Although the Vancouver Island results suggested no association at a l l , the Nelson seedlings i n replicates 2 - 4 gave results approaching significance. Perhaps the small sample size i s more at f a u l t than the b i o l o g i c a l relationship tested. The Seymour and Haney regressions were compared for s i m i l a r i t y of slope and intercept i n each case. They behaved the same i n replicate 1, but d i f f e r e d i n slope for the data from replicates 2 - 4 . This emphasizes again the strong influence of environment on the performance cf western hemlock seedlings. Several rows of seedlings were checked to determine more closely the relationship between f r o s t damage severity and bud development. Both damaged and undamaged seedlings were inspected to see whether or not a bud B - Rep 1 Nelson 7 -0.001000 Haney 40 -0.004912 Seymour 30 -0.004991 Vancouver Island 10 -0.001974 was v i s i b l e and to estimate i t s s i z e . No count or s t a t i s t i c a l analysis was made, consequently, the following i s based on an empirical assessment only. Damage occurred to trees with and without v i s i b l e terminal buds. Where damage was found, those plants without a v i s i b l e bud, or with small buds, were damaged more frequently and severely than the others. Thus, although there i s not a clear-cut r e l a t i o n s h i p between v i s i b l e bud and f r o s t hardiness, there i s an a s s o c i a t i o n i n the group showing damage. However, the expression of t h i s r e l a t i o n s h i p i s strongly influenced by environment, since one r e p l i c a t e here was damaged more severely than the others. Nursery vs. plug f r o s t damage The r e l a t i o n s h i p between f r o s t damage i n the nursery and i n the plug-grown material appears i n Figure 3 - 2 3 . The data were tested by regression a n a l y s i s f o r nursery r e p l i c a t e 1 and r e p l i c a t e s 2 - 4 pooled (Table 3-32 ). Table 3 - 3 2 Regression comparisons of f r o s t damage i n plug-grown stock and i n the nursery.Form: Plug f r o s t = a + b (Nursery f r o s t ) . Nursery Sample r e p l i -cate Slope: damage plug/ damage nursery Intercept damage plugs Stand-ard error Proba-2 b i l i t y r 1 87 47.92 2.06 10.88 0.186 * * 2 - 4 87 161.36 1.45 25.65 0.318 * * A highly s i g n i f i c a n t r e l a t i o n s h i p existed i n each case, but the data from r e p l i c a t e 2 - 4 gave a better 2 f i t than the others (r s of 0.318 and 0.186, r e s p e c t i v e l y ) . The regression from r e p l i c a t e s 2-4 gave a s i g n i f i -c a n tly higher slope than that from r e p l i c a t e 1, obviating the need to 162 80 \-70 60 50 40 30 20 10 •71 68 (19) . o o _ • CO O o c + O O . I- fiSDXOa—v,OA" o o o I I I I I 1 • 53 77 73 Legend: t> Nelson O Haney • Seymour + Caycuse X Paterson _L I 6 5 I JL JL 0 0.1 0.2 0.3 N u r s e r y - g r o w n F r o s t . d a m a g e r a t i n g Figure 3-23. Relationship between f r o s t damage percent of plug-grown seedlings and f r o s t damage rating of nursery-grown western hemlock seedlings, Reps 2-4 163 compare t h e i r intercepts. Yet the fact that a s i g n i f i c a n t relationship was found between the fr o s t damage incidence i n the plugs and i t s severity i n the nursery strengthens the assumption that the damage noted i n the plugs as described r e a l l y was fros t damage, and not due to s o i l desiccation or mold. While i t i s clear that there i s a relationship between fr o s t damage i n both nursery and plug-grown stock, the relationship i s not consistent for a l l of the three main c o l l e c t i o n s , those from Nelson, Haney and Seymour. Nursery damage rating i n the Nelson material was low throughout, although i t appeared to r i s e i n s i g n i f i c a n t l y for the c o l l e c t i o n from 5200 feet. Since t h i s l a s t value was based on only two f a m i l i e s , the re s u l t i s not necessarily r e l i a b l e . Both of the main Coastal c o l l e c t i o n areas, Haney and Seymour, showed an increased resistance to f r o s t damage with increased elevation of source above a certain lower a l t i t u d e (Figure 3-19). A break i n the damage trend occurred between 500 feet and 1300 feet for the Haney material, and between 1000 feet and 1900 feet for the Seymour material. Samples from intermediate elevations i n each case might have shown the breaking point of the trend to be d i f f e r e n t . The differences between the two areas are apparent for the f i r s t four points on each curve, where the collections are from reasonably comparable elevations. The Seymour coll e c t i o n s were damaged more heavily than t h e i r Haney counterparts i n each case: Haney material from 1300 feet was as fr o s t - r e s i s t a n t as Seymour material from 1900 feet. The same relationship was true for the Haney 1800-foot and the Seymour 2750-164 foot provenances. Thus, the trends of fr o s t s u s c e p t i b i l i t y are similar i n these two neighbouring areas, but the severity of damage suffered d i f f e r e d s i g n i f i c a n t l y , indicating a greater degree of f r o s t hardiness i n the Haney material under these conditions. The Vancouver Island c o l l e c t i o n s both displayed f r o s t hardiness char a c t e r i s t i c of material from higher elevations i n the Lower Mainland. Both were as hardy as the Haney 1300-foot and the Seymour 1900-foot co l l e c t i o n s . The consistent, inverse relationship between percentage budset and f r o s t damage for the Haney and Seymour seedlings, despite the higher damage i n replicate 1, indicates that there i s a strong general associar t i o n between bud development and f r o s t hardiness i n western hemlock provenances. Similar r e s u l t s were obtained i n Great B r i t a i n by Lines and Aldhous (1960, 1961),and for Douglas-fir by Campbell and Sorensen (1973). The s t a t i s t i c a l significance of any int e r - r e l a t i o n s h i p of these features for the Nelson and Vancouver Island c o l l e c t i o n s was weakened by the strong intra-provenance v a r i a t i o n , combined with the r e s t r i c t e d sample size. The significance of families within provenances indicates that i n t r a -provenance v a r i a b i l i t y i n f r o s t s u s c e p t i b i l i t y i s high and that values from more than one parent tree are needed to characterize a provenance's s u s c e p t i b i l i t y to f r o s t . Pooling seed from ind i v i d u a l parents w i l l r e s u l t i n a s i g n i f i c a n t loss of s e n s i t i v i t y to intra-provenance v a r i a b i l i t y i n behaviour. This should be remembered when any further work of t h i s nature i s planned with western hemlock. 165 From the general reduction i n standard error of damage (Appendix 8) control over the factors imparting f r o s t resistance, under the conditions i n the bare-root nursery, increases exponentially with seed source elevation. This can be seen best from the Haney data where the number of families studied varies only s l i g h t l y between a l t i t u d e s . The v a r i a b i l i t y within the lowest (100 to 500 feet) c o l l e c t i o n s i s the same; i t decreases to 2/3 of t h i s value by 1300 feet, but plunges to 1/10 the o r i g i n a l value i n the l a s t 500 feet. The results from Mount Seymour are not so clear-cut, p a r t l y because the 1000-foot population i s represented by only 3 fami l i e s , but the highest c o l l e c t i o n shows much less v a r i a b i l i t y than that from 1900 feet (850 feet lower): 0.0016 vs. 0.0079. These results suggest that the conditions requiring greater f r o s t resistance increase exponentially with a l t i t u d e , although inspection of the mean values revealed only a l i n e a r trend. The i r r e g u l a r i t y of the association between f r o s t damage and mean height of the plug-grown seedlings was quite d i f f e r e n t from the results of Nienstaedt (1958) with eastern hemlock provenances grown i n Wisconsin. In that study, nursery f r o s t damage, rated s i m i l a r l y to the system used here, was highly s i g n i f i c a n t l y correlated with the mean shoot extension obtained elsewhere i n a controlled-environment test. The main differences between that study and t h i s are that families were maintained separate here and the c o l l e c t i o n s came from a small part of the range, whereas they were combined into provenance col l e c t i o n s by Nienstaedt, and the col l e c t i o n s covered much of the range of eastern hemlock. Once again, the influence of nursery environment was strong. Despite 166 the agreement between damage rating i n the nursery and damage incidence i n the plugs, there was a s i g n i f i c a n t difference between nursery r e p l i -cates i n damage rating. The fact that a number of families escaped' damage i n replicate 1 when they sustained damage i n the rest suggests that environmental differences between rows i n a replicate were respon-s i b l e for some differences i n damage. Although i t seems u n l i k e l y i n the small amount of nursery space involved i n t h i s study, micro-climatic v a r i a t i o n during the conditions causing f r o s t damage may be p a r t l y responsible for the v a r i a t i o n found. Testing seedlings i n freezer chests with adjustable temperature control would permit more accurate deter-mination of the f r o s t s u s c e p t i b i l i t y of the seed sources i n t h i s t r i a l . Despite the evident importance of nursery environment to the damage observed, the clear reduction i n damage among high-altitude progeny from the Haney and Mount Seymour areas indicates that the populations sampled along these transects d i f f e r i n s u s c e p t i b i l i t y to f r o s t , and that these differences are genetically-determined. The creation of these differences w i l l have involved the elimination of those individuals from each popu-l a t i o n that were poorly suited to the climatic vagaries of the area. Yet nearly every population produced progeny that c l e a r l y were poorly suited to growing i n the nursery. Since there was no correlation with mean family height, non-genetic effects, such as weak seedlings from immature seed, should not be important. The "unsuited" types may have arisen following long-distance pollen transport (Lanner, 1966). If the pollen originated nearby, the v a r i a t i o n discovered probably i s a r e f l e c t i o n of the genetic v a r i a b i l i t y i n f r o s t s u s c e p t i b i l i t y existing within the l o c a l i t y . Genetic re-assortment during meiosis and f e r t i l i z a t i o n w i l l produce a broad array 167 of types each generation upon which selection by the environment can work. By the time sexual maturity i s reached, the number of frost-susceptible individuals could be reduced considerably, depending upon the persistence of the frost susceptibility found here and the timing and severity of frosts. The foregoing indicate that frost resistance i s under strong genetic control and i s of decided importance to populations growing at a l l ele-vations, but particularly those in high elevations. Definition of the nature of this genetic control w i l l require detailed studies involving controlled-pollinated seedlings and uniform environments in which to monitor response. 168 Summary of nursery studies The nursery studies conducted revealed many strong i n t e r s p e c i f i c differences between mountain and western hemlock, and some evidence of strong inter-and intra-population differences i n western hemlock. These differences were apparent i n a number of features that showed consistency of response, despite being drawn from material grown i n different locations and environments. Germination The feature of germination of most a n a l y t i c a l value was germinative energy. I t revealed the acceleration i n germinative speed produced by s t r a t i f i c a t i o n , as well as a strong difference i n speed between western and mountain hemlock. This difference, 6.2 days for u n s t r a t i f i e d seeds and 8 days a f t e r s t r a t i f i c a t i o n , can be used to test seed col l e c t i o n s suspected of being hybrids. Providing the seeds are f u l l y mature, hybrid seeds should emerge a mean of 3 days l a t e r than mountain hemlock, and 3 days e a r l i e r than western hemlock, assuming that germinative energy i s regulated by additive-effect genes. Analysis of germinative energy was not conducted on a l l seedlots, so differences between western hemlock provenances could not be tested. Indications of the genetic basis of germinative behaviour i n western hemlock are presented i n Chapter 4. Mutants The existence of chlorophyll-deficient "mutant" types i n most populations sampled suggests widespread s e l f i n g i n western hemlock, similar to that i n many coniferous species (Franklin, 1970). The i n -cidence probably w i l l vary with year as well as seed parent, as found here. Provided that environment has not affected th e i r expression, the occurrence of two mutant types i n the progeny of one seed parent indicates that more than one gene system i s responsible for the "mutants" i n t h i s case. The genetic nature of such mutants i n western hemlock i s indicated i n Chapter 4. The lack of albino types i n the mountain hemlock families may be only chance due to the r e s t r i c t e d sample available. In view of the commonness of chlorophyll-deficient types among coniferous genera (Franklin, 1970), i t i s probablg that similar germinants w i l l be found i n mountain hemlock also. Of the other mutant types studied, only the bluish type seems to occur i n any pattern. I t may r e f l e c t adaptation to high-altitude climates. Survival Although survival value i s an important attribute i n nature as well as forestry, the lack of v a r i a b i l i t y shown here by such young plants i s of l i t t l e importance. Survival i s meaningful only when assessed on s i t e s similar to those where the seed might be used (Silen, 1964b). The bare-root nursery and plug environments were made as uniform as possible to allow expression of features, such as aberrant seedling types, that might have been suppressed otherwise. Features of obvious importance to survival bud set and f r o s t resistance, were i d e n t i f i e d here. Since bud set i n p a r t i c u l a r seems affected by seed parent source elevation, i t should be followed i n outplanting t r i a l s . 170 Weak seedlings L i t t l e information was provided by analysing the proportion of weak seedlings i n the nursery beds. Mostly i t refected immature seed due to early cone c o l l e c t i o n . Lowest nursery survival at two years, corresponding with a s i g n i f i c a n t correlation between weak and unringed seedlings, was found i n the Nelson c o l l e c t i o n from 5200 feet. I t was collected when the seeds probably were immature. Otherwise no association with any important feature was found. Assignment of a seedling to a "weak" class i s a r b i t r a r y , and w i l l vary with the observer. In view of i t s random pattern of v a r i a t i o n i n western hemlock i t i s not worth including i n further studies unless X-ray plates of the seed samples are available f o r detailed examination and the seeds are sown equally spaced and reared under conditions more uniform than provided i n t h i s nursery study. Seed weight No connection was found between seed weight and height of plug-grown seedlings. This disagrees with the results of Buszewicz and Holmes (1960), also supported by Lines and Aldhous (1960), that t a l l e r seedlings come from provenances giving heavier seed. Since separate parent trees from a few provenances were studied here, rather than widely-separated provenances, the comparison i s not v a l i d . The results agree with those of Pieseh (1974) that there i s strong intra-population v a r i a b i l i t y i n seed weight and siz e . Screening seeds of one provenance to increase uniformity of seed size , as might be done to f a c i l i t a t e machine sowing i n a nursery, could r e s u l t i n the narrowing of the seed 171 parents remaining and therefore, of the genetic base of the seedling ; population. Genetic bases to western hemlock seed weight are explored i n Chapter 4. Height Although height i s limited by date of bud set, the mean height of the plug-grown seedlings was not correlated with bud set i n 1970, and reached s i g n i f i c a n t negative correlation with 1971 bud set only for the Haney material. Height generally was not associated with f r o s t damage in plugs or the nursery, although f r o s t damage did increase with height i n r e p l i c a t e 1, which showed the heaviest damage. Even there the assoc-i a t i o n occurred only i n the progeny of Haney trees. Therefore, seedling height of western hemlock families appears to be loosely associated with f r o s t resistance, as found by Hagner (1970 a & b) for Pinus s y l v e s t r i s L. and P_. contorta Dougl. No agreement i s found here with the suggestion by Lines and Aldhous (1960) that proximity to the Coast was important i n understanding the heights of western hemlock provenances. Where elevation of seed source i s comparable, (1800-2000 f e e t ) , the Int e r i o r progeny were t a l l e r than those from the Coast. The two main Coastal provenances, Haney and Seymour, dif f e r e d i n t h e i r trends of height with seed source elevation, whereas Haney and Nelson, an Int e r i o r source, had similar trends. However, there i s agreement with the B r i t i s h experience that height generally i s depressed as elevation of seed source r i s e s . 172 Height; diameter r a t i o Although mean diameter changed l i t t l e with elevation of seed sources, the height: diameter r a t i o did, following the trend of height decrease. This agrees with the re s u l t of the early European studies summarized by Langlet (1967), plus those of e.g. Callaham and Hasel (1961) and Hermann and Lavender (1968) with western coniferous species, that high-altitude populations are inherently stockier than t h e i r low-altitude counterparts. Yet there i s considerable v a r i a b i l i t y i n t h i s tendency within a population depending on parent tree, agreeing with the v a r i a b i l i t y within stands -found for other c h a r a c t e r i s t i c s . Chapter 4 presents more detailed consideration of western hemlock seedling height and diameter. Branch cha r a c t e r i s t i c s The branch characteristics measured on the plug-grown seedlings varied more between families within an area than between areas. There-fore, i t should be possible to choose seed parents i n a l o c a l i t y for th e i r tendency to produce fewer and shorter branches per unit of height without s a c r i f i c i n g s u i t a b i l i t y to the l o c a l climate. More precise indication of the bases of genetic control over western hemlock branch number and length w i l l be given i n Chapter 4. Bud burst Bud burst varied strongly within provenances, but was not useful i n detecting differences between the provenances used here. The winter temperatures may have s a t i s f i e d the c h i l l i n g requirements of a l l seed-l o t s f u l l y , removing differences i n phenology that might otherwise have 173 occurred (Worrall and Mergen 1967). B r i t i s h Forestry Commission ex-perience with western hemlock provenances i s that high-latitude and high-a l t i t u d e provenances flush e a r l i e r than those originating lower or more southerly, exposing them to damage by la t e spring f r o s t s . No such trend was found here, so bud burst threshold may not change with a l t i t u d e when the elevational difference i s so small. Bud burst seems very sensitive to temperatures and s o i l condition. Further studies of i t should be conducted i n a controlled environment, perhaps including d i f f e r e n t temperature and photoperiodic regimes to determine consistency of behaviour by provenances and families-.-. From the results here, strong differences i n time of bud burst can exi s t between families from the same stand, permitting selection of parents giving early or l a t e r bud burst for use i n planting s i t e s of d i f f e r e n t spring weather patterns, providing the testing conditions used r e f l e c t f i e l d performance adequately. Bud set Degree of bud set was the most useful measure of phenology observed, since i t revealed strong differences between mountain, and western hemlock and among western hemlock seed source elevations, and was associated with f r o s t resistance. I t was more useful i n the second year than the f i r s t , when s i g n i f i c a n t trends were discovered only l a t e i n the year. This may r e f l e c t greater ease of seeing buds on larger plants, or greater phenological " s t a b i l i t y " i n the older plants. Bud set was affected also by differences i n the s o i l environment, but less strongly i n the older plants. 174 Factors apparently affecting bud set were heat, and drought or l i g h t strength. Seedlings near the western sideboards set buds e a r l i e r i n both years, remaining short at age two years. Since a l l seed sources used originated i n a narrow l a t i t u d i n a l b e l t , differences i n photoperiodic stimulus to set buds should be less important than differences i n tem-perature threshold, i f they e x i s t . The factors c o n t r o l l i n g bud set by western hemlock populations from different altitudes within a region could best be studied i n a series of controlled environments combining differences i n temperature and photoperiod, as done by Robak and Magnesen (1970) for Norway spruce. The markedly e a r l i e r bud set by a l l mountain.hemlock families i n both years indicates that a strong difference exists between i t and western hemlock i n the factors or levels c o n t r o l l i n g bud set. This difference appears more profound than simply an ecotypic adaptation from a continuous population, and supports t h e i r d i s t i n c t i o n as separate species. Frost damage Frost damage served as a good screening of the other measures i n predicting the differences i n adaptation by the populations between and within areas. Although subject to modification by the nursery environ-ment, i t indicated c l e a r l y that populations of western hemlock from d i f f e r e n t elevations on two mountains (Haney and. Seymour) di f f e r e d s i g n i f i c a n t l y i n t h e i r f r o s t hardiness i n the nursery environment provided. Comparison of f r o s t damage trends with a l t i t u d e showed that the Haney and Mount Seymour co l l e c t i o n s gave sim i l a r responses to increase i n a l t i t u d e , unlike t h e i r trends of height growth. But they d i f f e r e d i n the l e v e l s of damage, which does agree with a s i m i l a r difference i n regression-predicted basal seedling height of the two areas. However, seedling height of plug-grown material was generally not associated with fro s t resistance, indicating that families from a p a r t i c u l a r elevation might be screened for fr o s t resistance without necessarily s a c r i f i c i n g juvenile height growth. Although f r o s t damage showed more s t a b i l i t y than most measures taken, f r o s t resistance should be assessed from tests conducted at known temperatures correlated with measures of plant a c t i v i t y : mor-phological ones of the sort used by Hagner (1970a & b), or physiological ones, such as e l e c t r i c a l impedance (van den Driessche, 1969 b). Environmental v a r i a b i l i t y One s t r i k i n g r e s u l t of the studies conducted i n the nursery i s the extreme s e n s i t i v i t y of western hemlock seedlings to environmental v a r i a b i l i t y . Soos (1961) also found strong mierosite influence on western hemlock seedling morphology. Despite establishing the nursery study i n small beds containing specially-mixed s o i l , s u f f i c i e n t v a r i a -b i l i t y existed i n the s o i l to affect bud burst, bud set i n both years, and f r o s t damage. The pattern of v a r i a t i o n i n the beds established i n 1970 was not altered s i g n i f i c a n t l y by f e r t i l i z e r placement i n 1971, so that the more " f e r t i l e " areas responded s i m i l a r l y to the obviously less f e r t i l e areas. Any further studies of a similar nature on western : hemlock seedlings should be conducted under more uniform conditions than available here. Further discussion on t h i s point i s presented i n Chapter 4. 176 3.6 Inter - and i n t r a - s p e c i f i c inferences from population studies. 3.61 In t e r - s p e c i f i c comparisons The res u l t s of these studies show a number of marked differences between western and mountain hemlocks. Their cones d i f f e r consistently i n scales per centimeter, scale length : width r a t i o , the r e l a t i v e lengths of bract and scale, and bract shape. Cone length i s less re-l i a b l e , although the mean values found agree with those of Taylor (1972). Mountain hemlock seeds germinated faster, the seedlings formed and broke buds e a r l i e r , and exhibited much higher f r o s t resistance than the western hemlock provenance collected closest to i t , or that collected highest of a l l i n the study: Nelson, 5200 feet. Also, i t s pattern of shoot extension i s of a series of flushes interrupted by the formation of a temporary bud, whereas that of western hemlock apparently i s of un-interrupted elongation of a single axis u n t i l the "resting" bud forms. The degree and extent of these morphological and phenological differences between the species, even when only the western hemlock co l l e c t i o n s from the highest elevations are considered, are such that more than a l t i t u d i n a l or ecotypic differences alone apparently are responsible. These seem to be characteristics of di f f e r e n t species. Although study of the cone morphology of a population using the features outlined w i l l permit accurate assignment of i t to the correct species, testing seedlings of a population giving intermediate cone values against samples from v e r i f i e d c o l l e c t i o n s of each species for r e l a t i v e germinative rate, time of bud burst, nature of height growth (continuous vs. intermittent), and time of bud set w i l l permit f a i r l y 177 s e n s i t i v e assessment of i t s p o s i t i o n r e l a t i v e to the pure species. Assuming that the phenomena mentioned are governed by several pro-cesses,- themselves regulated by many genes, the intermediate popu-l a t i o n should f a l l between the "pure" species i n these features. 3.62 I n t r a - s p e c i f i c inferences for western hemlock Inter-population v a r i a t i o n patterns A l i m i t e d t e s t of inter-population cone v a r i a b i l i t y indicated no consistent pattern between Coastal patterns and I n t e r i o r populations from comparable elevations. Inter-population differences i n seedling behaviour along a l t i t u d -i n a l gradients were found f o r height, height: diameter r a t i o , number of branches, bud set and f r o s t damage (Table 3-33 ). With some exceptions, h i g h - a l t i t u d e seedlings were shorter, s t o c k i e r , produced fewer branches, set buds e a r l i e r and exhibited greater f r o s t r e s i s -tance. From low to high seed sources on a mountain, the v a r i a t i o n was generally c l i n a l . S t a t i s t i c a l t e s t s of the western hemlock population data are. summarized i n Table 3-33 . 178 Table 3-33 Summary of seedling studies on open-pollinated western hemlock: resume of s t a t i s t i c a l tests, A. Analysis of variance A N O V A T E R M Feature Area Elevation in.area Seed parent i n elevation Bud set Dec. 12, 1970 Bud set Oct. 20, 1971 Frost damage, nursery reps 2-4 S I G N I F I C A N C E Height Height/Diameter Adjusted longest branch *** N.S. Adjusted branch number Bud burst 1971 N.S. N.S. N.S, Frost damage, nursery rep. 1 N.S. N.S. 179 Table 3- 33 (continued) B. Linear regressions F e a t u r e s S e e d s o u r c e a r e a s Dependent Independent Nelson Haney Seymour Vancouver Island A l l pooled Correlation c o e f f i c i e n t and significance Height Elevation -0.476** -0.264** -0.200** Longest branch (adj) Elevation -0.061 -0.191** -0.088 Bud set Dec. 12, 1970 Elevation 0.633 * 0.344* 0.268 Bud set Oct. 22, 1971 Elevation 0.633* 0.622** 0.746** Height Bud set Oct. 20, 1971 -0.670 -0.466** -0.286 -0.358 -0.319** Frost damage Reps 2-4 Bud set Oct. 20, 1971 -0.660 -0.520** -0.461* -0.095 -0.488** Frost damage Rep. 1 Bud set Oct. 20, 1971 -0.338 -0.470** -0.634** -0.441 Frost damage Reps 2-4 Height -0.675 -0.079 -0.054 -0.197 -0.013 Frost damage Rep. 1 Height -0.475 -0.515** -0.206 -0.656* -0.316** Frost damage Reps 2-4 Elevation -0.210 -0.254** -0.283** Frost damage plugs Frost dam-age Rep. 1 0.431** Frost damage plugs Frost dam-age Reps 2-4 0.563** 180 C l i n a l v a r i a t i o n i s inferred from the presence of l i n e a r regressions of seedling behaviour on a l t i t u d e of source. Calculation of a regression assumes a l i n e a r relationship, but none was calculated here before a p l o t of the data had been inspected. Although s i g n i f i c a n t regressions on elevation were found for seedling height and f r o s t damage, the strong-est association with elevation was found for bud set i n 1971 (Table 3-33). This indicates a strong adaptive role for bud set i n the l o c a l i t i e s sampled. Since the association between bud set and elevation of source i s p o s i t i v e , bud set l i k e l y i s important i n c o n t r o l l i n g height growth and i n i t i a t i n g dormancy. But the weaker associations of height and f r o s t resistance with bud set indicate that the relationships of these features to bud set are hot d i r e c t . With the widespread occurrence of the c l i n a l v a r i a t i o n found i n western hemlock here, some feature of the environment that varies simi-l a r l y should be i d e n t i f i a b l e as the p r i n c i p a l selective force. Aspect of most co l l e c t i o n s i n the transects was the same: south, so ecotypic differentation of the kind found by Squillace and Bingham (1957) and Hermann and Lavender (1968) should not be involved. P r e c i p i t a t i o n gen-e r a l l y increases with a l t i t u d e , p a r t i c u l a r l y near the ocean. This should favor growth at higher a l t i t u d e s , not r e s t r i c t i t , as found. However, i n -creased leaching of nutrients with the higher r a i n f a l l may cause growth to be l i m i t e d by available n u t r i t i o n . Percent of p r e c i p i t a t i o n as snow i n -creases strongly with elevation (Table 3-2 ) i n a l i n e a r or c u r v i l i n e a r fashion. Length of the growing period varies s i m i l a r l y (Table 3-4 ), and l i k e l y i s associated with the former feature. Snow load was i d e n t i f i e d by Thornburgh (1969) as the major agent eliminating western hemlock seed-181 l i n g s from the upper elevations of the true fir-hemlock forests i n the Washington Cascade mountains. However, sampling here covered more than the upper margin of western hemlock's elevational d i s t r i b u t i o n . E v i -dence of adaptation to difference i n elevation was found throughout the zones sampled, rather than at the top only, where snow load might i n -crease exponentially. This points strongly to the p r i n c i p a l selective force being decreasing growing period, defined here as minimum f r o s t -free period (Table 3- 4 ) . Although high-altitude populations may be able to function s a t i s -f a c t o r i l y at lower temperatures than t h e i r low-altitude counterparts (Fryer et a l . , 1971), they w i l l be i n temperatures prohibiting growth a greater proportion of the time than populations lower on the slope. Thus, growing season length w i l l be a feature of.great importance to the population at each a l t i t u d e . And because of the "screening" effect on gene frequencies by infrequent, drastic fluctuations i n growing season length, minimum growing period l i k e l y w i l l be more indicati v e of the inherent growth rhythm of l o c a l plant populations than average length of the growing period. Despite the evident importance of seed source a l t i t u d e to proven-ance behaviour i n a number of features, agreement with the clim a t i c data assembled i s not good. The best regressions with a l t i t u d e were obtained for bud set on October 20, 1971 (Table (3-27 ). A l l of the main sampling areas gave s i g n i f i c a n t regression slopes and intercepts, agreeing with the s i m i l a r i t y i n trends of minimum fr o s t - f r e e period presented i n Table 3-4 . S i m i l a r l y , the Nelson 2000 foot c o l l e c t i o n showed bud set degree the same as Coastal material from comparable elevations, 182 r e f l e c t i n g the s i m i l a r i t y of autumn weather presented i n Table 3- 4 . Nursery damage the following February showed d i f f e r e n t patterns. Seymour was damaged more heavily than Haney progeny for a l l comparable elevations. This suggests that the two areas d i f f e r i n t h e i r rates of progress toward f r o s t hardiness despite t h e i r s i m i l a r i t y of bud formation. Haney material seems adjusted to grow i n an area with e a r l i e r or harsher f r o s t s than does that from Seymour. The values i n Table 3-4 suggest that the two areas share a common f a l l climate, whereas the response to f r o s t indicates d i f f e r e n t l y . Thus, date of f i r s t f a l l f r o s t may not be the main force of selection for f r o s t re-sistance with increasing a l t i t u d e of population, but i t seems to be important to selection for early bud set. Perhaps f r o s t resistance i s more conditioned by weather pattern i n each year, than by minimum frost-free period, A combination of frost-free period and the f r e -quency and severity of polar outbreaks might be more indicati v e of the adaptions of these Coastal populations to f r o s t stress, but the climatic data available now are inadequate to explain more than the pattern of bud setting. Intra-population v a r i a t i o n Although there were decided differences between populations of western hemlock from different elevations i n three separate areas, there was always strong intra-population difference. For both cones (Table 3-7 ) and seedlings (Table 3-33A), intra-population v a r i a b i l i t y was as strong or stronger than that between populations. The only nursery r e s u l t s for which s i g n i f i c a n t intra-population v a r i a t i o n was 183 not found were of f r o s t damage i n rep l i c a t e .1. This indicates that differences between cone parents can be as great as those between a l t i t u d i n a l provenances, or between those from dif f e r e n t locations i n the same l a t i t u d i n a l and elevational b e l t . This heterogeneity might be the re s u l t of-the infusion of genes into each of the c o l l e c t i o n areas from locations characterized by sharply d i f f e r e n t environments. Bernsten's (1955) study of western hemlock seed dispersal showed that upslope movement was possible i n view of the time seed could remain a l o f t . This has been corroborated by Thompson (1968) i n the Interior of B r i t i s h Columbia. Appropriate upslope winds are common during warm f a l l weather, when a i r i s heated over exposed slopes. But Hetherington's (1965) observation that most seed was released during cool, dry winter periods when the wind came from the north or east tends to reduce the li k e l i h o o d that upslope seed movement i s responsible for the heterogeneity found i n a l l the stands sampled. Important upslope winds, are more l i k e l y to occur during the period of pollen release. I f trees at higher elevations overlap i n time of anthesis and pollen r e c e p t i v i t y with those downslope, as found by G r i f f i t h (1968) for Douglas-fir i n a more r e s t r i c t e d area, the effective area of p o l l i n a t i o n by a tree could be extensive. Sakai and Park (1971) found s i g n i f i c a n t differences i n enzyme banding pattern between natural populations of Cryptomeria japonica D. Don. separated as l i t t l e as 700 meters horizontally at the same approximate elevation i n a mountainous area. They concluded that pollen migration occurred over only short distances. Strand (1957) found that elevation within a small area 184 provided marked delay i n strobilus receptively and anthesis, decreasing the chance of strong impact on the general seedcrop by upslope pollen movement. However, intra-stand differences i n time of pollen r e c e p t i v i t y could make the small amount of pollen moving upslope an important factor i n the behaviour of the progeny of the early-flowering trees, and t h e i r pollen, i n turn, could f e r t i l i z e the late-flowering trees down-slope. Al t e r n a t i v e l y , the strong heterogeneity within stands could r e f l e c t genetic heterogeneity. This could stem i n part from the nature of the habitats encountered by forest trees; high v a r i a b i l i t y i n seedbed and microclimate, and strong competition for n u t r i t i o n and growing space during seedling establishment. These forces, plus the customary longevity of forest trees, combined to make most tree species outbreeding and heterozygous, rather than inbreeding and homozygous (Rehfeldt and Lester, 1969, Libby et al., 1969) . As discussed before (Chapter 2), western hemlock reproduces from seed and occupies a wide range of lat i t u d e and a l t i t u d e , indicating that i t possesses adaptations to many climates and environments within them. Thus, i t probably w i l l be outbreeding and heterozygous. Depending on the number of genes segregating, meiosis and recombination w i l l produce a broad range of seedling types from which environmental stresses w i l l eliminate those unable to reach reproductive age and contribute to the next generation. Intra-population v a r i a b i l i t y was s i g n i f i c a n t i n studies of cone morphology, seedling height, and height/diameter r a t i o , adjusted branch length, branch number, bud burst both years, bud set and fr o s t hardiness. A l l of these but cone morphology involve regulation of elongation, mostly 185 by the main axis , and l i k e l y are controlled by similar physiological systems, thereby involving only a small proportion of the genotype. However, considering the complexity of the processes involved i n receiving s t i m u l i that i n i t i a t e growth and regulate i t , then terminate i t and develop f r o s t hardiness, i t i s l i k e l y that a number of gene complexes are involved. This suggests that the differences between families within a stand r e f l e c t differences between thei r parental genotypes. Similar differences l i k e l y occur between populations distributed along an environmental gradient, such as a mountain slope. Although i t i s probable that gene flow along an a l t i t u d i n a l gradient i s r e s t r i c t e d by limited seed and pollen f l i g h t , and by upslope delay i n strobilus r e c e p t i v i t y , some must occur where aspect i s the same, as the Mount Seymour and Haney co l l e c t i o n s . Several generations of selection w i l l be required to produce a population closely adapted to a p a r t i c u l a r l o c a l i t y , yet v a r i a b i l i t y within i t should s t i l l be pronounced because of the v a r i a b i l i t y of s i t e s , the occasional infusion of genes by outside pollen, and because of the latent heterozygosity i n most tree populations. This appears to be the case with the western hemlock populations studied: adaptation to l o c a l conditions has occurred with elevation, but intra-population v a r i a b i l i t y i s high. Western hemlock seems to have responded to environmental change along a l t i t u d i n a l gradients s i m i l a r l y to Douglas-fir (Hermann and Lavender, 1968), ponderosa pine (Callaham and Liddicoet, 1961), and balsam f i r (Fryer et a l . , 1971). The degree of adaptation seems quite sensitive between areas, since differences 186 i n bud set l e v e l and f r o s t resistance were found between slopes only 17 miles apart (Haney and Seymour). Adaptation within area was f a i r l y regular i n the Haney material, but ir r e g u l a r i n the Mount Seymour col l e c t i o n s . This possibly r e f l e c t s differences i n the environmental pressure on populations i n the two areas, f r o s t resistance perhaps being at a higher premium near Haney than on Mount Seymour. Yet the marked v a r i a b i l i t y within stands supports the idea that western hemlock populations are highly heterozygous, and therefore that in d i v i d u a l trees l i k e l y are outbreeding. Further evidence on t h i s question w i l l be presented i n Chapter 4. The commonness of tree-to-tree v a r i a b i l i t y i n the t r a i t s i n v e s t i -gated could permit a tree breeder to select strains for p a r t i c u l a r purposes from a single stand. Provided he has adequate information on the s i m i l a r i t y i n climates of the seed source and the planting area, a breeder can choose parents to give differences i n seedling height, branch number and length, bud set and f r o s t hardiness, and perhaps bud burst threshold. A l l of these features can be assessed i n seedling material, but environmental uniformity i s imperative for a meaningful screening, since young western hemlock seedlings seem very sensitive to s l i g h t differences i n environment. More detailed investigation of the genetic bases i n western hemlock of the morphological features used here (height," diameter, branch length, etc.) w i l l be conducted on the material produced by controlled crossing and grown i n a more closely-regulated environment (Chapter 4). 187 Chapter 4 INTER-SPECIFIC AND INTRA-SPECIFIC CROSSES WITH WESTERN AND MOUNTAIN HEMLOCKS When i n i t i a t i n g a breeding program i n an organism, the breeder i s dependent on s u f f i c i e n t v a r i a t i o n existing i n the t r a i t ( s ) chosen to permit selection of desireable individuals f a i r l y e a s i l y , and on control of the t r a i t being p a r t l y genetic, permitting improvement by breeding and selection from the progeny of suitable individuals. Selection i s most f r u i t f u l where there i s considerable genetic v a r i a b i l i t y i n the t r a i t desired (Wright, 1962). Where more v a r i a b i l i t y i s desired, i t may be generated by, for example, i r r a d i a t i o n or hybridiza-t i o n . Consequently, hybridization i s a common feature of many tree-improvement programs at some point i n the i r development. Furthermore, information on the mode of reproduction and. genetic control over the t r a i t s desired i s needed to plan the breeding program. The li m i t e d knowledge of the relationship between mountain and western hemlocks, and of the genetic v a r i a b i l i t y of the l a t t e r were summarized i n Chapter 2. Studies involving controlled p o l l i n a t i o n s of ind i v i d u a l mountain hemlocks and western hemlocks were conducted i n 1970 to elucidate the genetic control over the features used and some of the inferences drawn i n Chapter 3, s p e c i f i c a l l y : the a b i l i t y of the two species to be hybridized a r t i f i c i a l l y , the incidence of apomixis i n each species, the a b i l i t y of western hemlock to be selfed, the importance of cone-pollen parent com-bination to the seed y i e l d and short-term performance of western hemlock seedlings, the genetic bases to the seedling features analysed i n the western hemlock population studies, and t h e i r relationships to t o t a l and 188 component seedling weights. 4.1 I n t e r s p e c i f i c crossing t r i a l with western and mountain hemlocks 4.11 Purpose and methods In an attempt to test the closeness of the relationship between western and mountain hemlock, controlled p o l l i n a t i o n s using western hemlock pollen were conducted on three mountain hemlock trees. Parent trees The area on Mount Seymour from which open-pollinated mountain hemlock seed was collected i n 1968 was checked i n May to locate trees carrying cone buds. To be useful for t h i s t r i a l , a tree had to be climbable and be producing many s t r o b i l i d i s t r i b u t e d over the crown so that a number of i s o l a t i o n bags could be attached. Three trees, were used. A l l were close to the parking l o t (Figure 4-1 ), and one (Tree A) had been i n -cluded i n the 1968 c o l l e c t i o n as tree 109. Isolations were conducted on May 21, when cone buds could be i d e n t i -f i e d . Foliage and any fine branches were removed c a r e f u l l y from the main branch, at the point of attachment of the bag. The branch then was wrapped with a wad of non-absorbent cotton wool which was held i n place with tape. I f foliage ahead of the s t r o b i l i was too long to f i t into the i s o -l a t i o n bag e a s i l y or might inter f e r e with p o l l i n a t i o n , i t was trimmed back, but leaving the t i p s s l i g h t l y longer than the s t r o b i l i where possible to reduce the l i k e l i h o o d that they would rub against the bag and suffer damage. Isolation bags were made of non-woven terylene material, were 9 x 12 or 9 x 15 inches i n size and included a 6 x 3-inch observation 189 Figure, 4-1 Location of mountain hemlock trees "A", "B", and "C" at Mount Seymour s k i h i l l used i n the hybridization attempt with western hemlock 190 window of polyvinyl chloride jtoward the bottom on one face only. These bags are widely used at present for hybridization work with forest trees as they are sturdy and water-repellant, but allow a i r and moisture ex-change, reducing the chance of heat damage or mold build-up. The bags were placed around the branch and attached with wire,"Twist-ems" around the cotton r o l l . The observation window was turned downward and away from the sun. These were checked periodically and turned where necessary to prevent the sun shining through the window and causing possible overheating. The number of s t r o b i l i isolated in each bag and the date of isolation were written on a tag attached to the branch near the bag. Pollen collection and extraction Branches bearing microstrobili (pollen cones) were cut from 3 western hemlock trees at Totem Park, U.B.C,, when the pollen s t r o b i l i were appearing between the bud scales and i t appeared anthesis was imminent. They were placed in separate, but adjacent rooms in the G.S. Allen Forest Genetics-Tree Seed Laboratory in the Macmillan Building, U.B.C. The branch butts were cut again and placed in jars of tap water to maintain branch moisture u n t i l the pollen was shed. Pollen was collected from paper sheets below each spray of branches, screened to remove d i r t , and stored in vials at 1-2° C u n t i l needed. To reduce the chance that the proximity to anthesis had been misjudged, two collections of branches, one to three days apart, were made from each tree. Portions from each collection were mixed be-fore pollinations were commenced. This mix was used for a l l pollinations. 191 P o l l i n a t i o n The number of s t r o b i l i available for crosses on each tree were assigned, as far as possible, to provide a nearly equal number of s t r o b i l i and bags for each cross. An attempt was made to d i s t r i b u t e the bags assigned to a p a r t i c u l a r cross throughout the crown to reduce the chance of biasing the resu l t s i f intra-tree v a r i a t i o n i n seed charac-t e r i s t i c s existed, and to reduce the r i s k of accidental loss of a l l bags of one type. P o l l i n a t i o n was conducted using a De V i l b i s s atomizer with a capacity of 10 cc. For dry po l l i n a t i o n s i t was f i l l e d completely. For wet p o l l i n a -tions (Allen and S z i k l a i , 1962) approximately 2 cc of pollen was mixed with d i s t i l l e d water immediately before application. One atomizer was used throughout the t r i a l for the pollen from each tree. Between v i s i t s to the trees, the dry pollen was stored i n the cold room at the same temperature as described above, but not under desiccation. The bags on each tree were checked frequently to follow the degree of opening of the s t r o b i l i . P o l l i n a t i o n was conducted when the majority of v i s i b l e s t r o b i l i had reflexed bracts. Pollen was sprayed through each of three small ports on the same face of the i s o l a t i o n bag as the viewing window. As far as possible i t was sprayed on a l l s t r o b i l i from more than one d i r e c t i o n . Dry pollen was sprayed u n t i l i t was v i s i b l e on the bag walls. During wet p o l l i n a t i o n the atomizer was shaken p e r i o d i c a l l y to prevent pollen from s e t t l i n g to the bottom. P o l l i n a t i o n ports were covered with masking or e l e c t r i c i a n ' s tape before and after p o l l i n a t i o n . P o l l i n a t i o n was carried out on tree "C" on May 24 and repeated on May 27. On trees "A" and "B" i t was applied on June 2 and not repeated 192 because the s t r o b i l i appeared f u l l y open, judging by the development of the s t r o b i l i on tree "C". The i s o l a t i o n s and p o l l i n a t i o n s conducted appear i n Table 4- 1. Table 4- 1 Cone Parent (Mountain) hemlock Mountain hemlock-western hemlock hybridization t r i a l , Number of s t r o b i l i by p o l l i n a t i o n method. P o l l e n p a r e n t None Bags Strob. Western hemlock A D E Bags Strob. Bags Strob. Bags Strob. A Dry Wet B Dry 25 23 2 12 1 6 2 16 11 12 17 C Dry 60 68 52 59 A l l bags were l e f t i n place for the entire season to prevent cone loss to squirrels and birds. The apparent maturity of the cones was f o l -lowed during September and October i n order to delay picking u n t i l seed release was imminent or had begun. At c o l l e c t i o n , each branch was cut below the bag and transported to the laboratory with the bag s t i l l i n place. There the number of mature cones per bag was counted and any loose seed was placed i n a labelled envelope and stored i n a cold room at 1 ° + 2 ° C. A l l cones were allowed to open at room temperature before the seed was removed by hand. Where possible, open-pollinated "check" cones were collected near each bag. Seed were cleaned by hand. A l l seeds from the bagged cones were examined v i a a flouroscopic screen on a "Softex" X-ray machine to determine 193 whether any " f i l l e d " seed had resulted. Any such seeds were separated from the rest by hand and photographed under the X-ray c e l l to permit more accurate determination of the contents. Seeds were placed i n d i v i d u a l l y i n c a v i t i e s on r i c e paper-backed cardboard sheets. The exposure and development schedule found most satisfactory was: distance 35 cm., power 125 kV and 6 mA, exposure time 45 seconds at small focus. Kodak type "M" i n d u s t r i a l f i l m was developed for Ah minutes. This produced a sharp, dark image. 4.12 Results Phenology The receptive p i s t i l l a t e s t r o b i l i of mountain hemlock are erect, later becoming pendent. In t h i s respect they d i f f e r from those of western hemlock, which were always pendent on the trees used i n the other study. Also, the female s t r o b i l i of mountain hemlock were generally well separated from the male s t r o b i l i , which were borne closer to the tree stem. Protogyny was pronounced on each tree. At the time of hand p o l l i n a -t i o n on each tree the female s t r o b i l i both i n and out of the bags were f u l l y receptive, yet pollen release was not seen from any other part of the tree. The pollen s t r o b i l i seen during the controlled p o l l i n a t i o n s appeared close to anthesis, indicating that some overlap of pollen release and female r e c e p t i v i t y might occur. As noted above, bagging had not ad-vanced the receptive time of the female s t r o b i l i over those exposed to the ambient temperatures. . Cone harvest A l l cones were collected on October 13 and 15, st a r t i n g with tree 194 A. Trees B and C were collected on October 15 only. Some unbagged cones were releasing seeds on those days, as were some of those i n bags. The number of cones harvested by cone and pollen combination appear i n Table 4-2. No effect from wet p o l l i n a t i o n was found on cone survival (data not shown). Table 4-2 Mountain hemlock - western hemlock hybridization t r i a l . Number of s t r o b i l i isolated and cones harvested by cross. Cone P o l l e n p a r e n t Parent W e s t e r n h e m l o c k (Mountain None A D E hemlock) Strob. Cones Strob. Cones Strob. Cones Strob. Cones A 25 22 - 18 16 23 19 B 23 19 - 16 13 17 16 C 60 54 68 60 52 46 59 50 The abortion loss of s t r o b i l i between i s o l a t i o n and harvest was low, averaging 12.5 percent. Some s t r o b i l i apparently were aborting when pollen was applied, and some were knocked off by the motion of the bags. One bag on tree A containing 5 s t r o b i l i was not pollinated because of the apparent abortion of the s t r o b i l i a f ter bagging. This was the only such case found. A few bags were l o s t during the season, but i n no pattern associated with pollen parent. Seed y i e l d Seeds recovered from each bag were separated into "empty" and " f i l l e d " portions under the X-ray fluoroscopic c e l l . No count was kept of seeds collected. Some seeds from both unpollinated "control" bags and from bags to which western hemlock pollen were applied gave dark images on the X-ray flouroscopic screen. A l l such seeds were kept and photographed using the X-ray exposure schedule ou t l i n e d before. Examples of the plates obtained appear i n Figures 4- 2 and 4-3 . For comparison, Figure 4- 4 presents a p l a t e of " f i l l e d " seed extracted from open-pollinated cones of tree A. Only the open-pollinated "check" cones produced normal f i l l e d seed. No germination t e s t of these seeds was run to see i f they were v i a b l e , but previous experience with mountain hemlock seedlots was that most seeds g i v i n g s i m i l a r X-ray images germinated w e l l . No d i f f e r e n c e i n image e x i s t s between the unpollinated and cross-p o l l i n a t e d seeds. Each produced some seeds showing what appears to be a collapsed mass of female gametophytic t i s s u e . When these seeds were cut open they contained a loose, collapsed mass of t i s s u e surrounded by a tough, whitish skin. On the basis of past experience these could not germinate, so no germination t e s t was run with these seeds. ^ £ Q Q $ Figure 4-2 X-ray photo of unpollinated mountain hemlock seed. Tree B. 196 Figure 4-3. X-ray of c r o s s - p o l l i n a t e d mountain hemlock seed. Tree C x western hemlock D. 197 Y0M 1 00 e&e Figure 4-4 X-ray photo of wind-pollinated seed of mountain hemlock. Tree A. 4.13 Discussion and conclusions Since no microscopic examination of the p o l l e n grains or structures i n the ovule were made, no information i s a v a i l a b l e on the normalcy of the development of the p o l l e n tube and the fate of any zygote formed, such as obtained f o r Douglas-fir by Orr-Ewing (1957a), by Hagman and 198 Mikkola (1963) and Kriebel (1967) for soft pines. Kriebel (1967) found that embryo gametophytic breakdown following i n t e r s p e c i f i c crosses.in the Strobus section of Pinus occurred after f e r t i l i z a t i o n , whereas i t occurred during pollen tube elongation i n several sections of the hard pines. The fact that the same X-ray image was obtained from seed from both unpollinated control bags and those to which western hemlock pollen was applied indicates that apomixis does not occur i n mountain hemlock and that the western hemlock pollen did not stimulate develop-ment within the seedcoat beyond that occurring naturally before abortion took place. During June of 1970, temperatures at Vancouver International Airport, 15 miles south-west of Mount Seymour, were higher than normal, hours of bright sunshine were up 20% and p r e c i p i t a t i o n was 18% lower than normal. S i m i l a r l y warm and dry conditions were indicated at stations nearer the study area. Perhaps retaining the i s o l a t i o n bags a l l season meant that excessive temperatures occurred i n the bags, i n -h i b i t i n g pollen development or damaging the developing embryos i f f e r t i l i z a t i o n had taken place. This possible source of embryo abortion could not be checked because no mountain hemlock was available. The r e s u l t s are the same for t h i s t r i a l as that conducted at C o r v a l l i s : no f i l l e d seed (Sorensen, 1974). On the basis of these te s t s , western and mountain hemlocks do not seem to be closely related and t h e i r hybridization as t r i e d here does not appear to permit a broad-ening of the gene pool for western hemlock improvement. 199 Van Campo-Duplan and Gaussen (1948) considered the pollen parent of the hybrid Tsuga x j e f f r e y i to be mountain hemlock because of the gametophytic control of pollen exine, including the "sacs". Since the hybrid lacks them, the seed parent should be western hemlock. To my knowledge the cross has not been t r i e d i n t h i s d i r e c t i o n (de Ferr£, 1973). This might be the next step i n attempts to hybridize these species, with the inclu s i o n of some western x western hemlock crosses to check on the pollen ef f i c a c y and application technique. 4.2 In t r a - s p e c i f i c crossing t r i a l with western hemlock 4.21 Purpose and methods A small d i a l l e l crossing program was conducted at U.B.C. to determine: y i e l d of seed and seedlings per cone from p o l l i n a t i o n type (selfed, out-crossed or open p o l l i n a t e d ) , the frequency of apomictic development of viable seed, and the influence of parental combination on the seed y e i l d and short-term growth of seedlings. A l l data were analysed to detect evidence of genetic and non-genetic control over each feature, although i t was realized that such small d i a l l e l s might provide unreliable r e s u l t s for the species (Yates, 1947, Hayman, 1954, Kempthorne, 1956). Parent trees Three naturally-occurring western hemlock trees at Totem Park on the campus of the University of B r i t i s h Columbia were selected. They were suitable for t h i s crossing t r i a l because each had a f u l l , deep, open-grown crown, bore a good crop of s t r o b i l i , and had borne cones recently. Details 200 of the trees' ages and dimensions are presented i n Table 4- 3. Figure 4- 5 shows th e i r locations. Table 4- 3 Dimensions of western hemlock trees used i n 1970 crossing t r i a l Approx. Total Crown Total D.B.H. height Class Tree age, years i n . f t .  A 56 20.5 97 Dominant D 54 15.2 59 Co-dominant E 49 18.1 90 , Co-dominant Isolation technique Isolations were conducted i n la t e March, when cone buds could be i d e n t i f i e d . Isolation techniques were described i n Section 4.11. Where a bag was to enclose a "shaker" s e l f - p o l l i n a t e d t r i a l , a l l pollen s t r o b i l i present were retained. Otherwise a l l were removed c a r e f u l l y to permit cross-pollination or to establish a non-pollinated "control" as a check on i s o l a t i o n efficacy. Pollen c o l l e c t i o n and extraction Branches bearing m i c r o s t r o b i l i (pollen conelets) were cut from each tree when the pollen s t r o b i l i were appearing between the bud scales and i t appeared anthesis was imminent. Extraction and storage was the same as presented i n Section 4.11. To reduce the chance that the proximity to anthesis had been misjudged, two co l l e c t i o n s of branches, one to three days apart, were made from each tree. Portions from each c o l l e c t i o n were mixed before p o l l i n a t i o n s were commenced. This mix was used for a l l p o l l i n a t i o n s . 201 Scale: 1 inch = 40 feet (approx.) Figure4-5. Locations of western hemlock trees used i n the complete d i a l l e l crossing program, Totem Park, U.B.C. 202 P o l l i n a t i o n As for the hybridization t r i a l , the number of macrostrobili (seed conelets) available for crosses on each tree were assigned to provide a nearly equal number of s t r o b i l i and bags for each cross. The number assigned to s e l f - p o l l i n a t i o n was high i n anticipation of reduced seed y i e l d . As before, the bags assigned to a p a r t i c u l a r cross were d i s t r i -buted throughout the crown. P o l l i n a t i o n was conducted using a De V i l b i s s atomizer with a capacity of 10 cc. as described previously. An attempt was made to use a hypodermic syringe and needle to spray pollen onto the s t r o b i l i , but the pollen plugged up the needle quickly and t h i s device was abandoned. The bags on each tree were checked d a i l y to determine the degree of opening of the s t r o b i l i . P o l l i n a t i o n was conducted when the majority of v i s i b l e s t r o b i l i had reflexed scales and apparently were receptive (Stanlake and Owens, 1974). In "shaker" s e l f - p o l l i n a t e d bags, pollen was disseminated by shaking the bag u n t i l pollen could be seen inside. As for the other bags, they were shaken when the s t r o b i l i appeared receptive. Isolations and p o l l i n a t i o n s conducted appear i n Appendix 10 A l l p o l l i n a t i o n s on each tree were repeated on two separate days to i n -crease the chance of successful p o l l i n a t i o n . The dates involved for each tree were: Tree A: A p r i l 1 and 3; Tree D: A p r i l 2 and 3; Tree E: A p r i l 3 and 6. Observations made on the s t r o b i l i of each tree on the days of p o l l i n a t i o n are presented i n Appendix 11 . Following p o l l i n a t i o n , periodic checks of the bags were made during A p r i l and May to see whether any obvious aberrations had occurred among the bagged cones vs. those 203 open-pollinated. The dates involved and the notes taken appear i n Appendix 11 . A l l estimates of cone size are based on a v i s u a l estimate. Aborted conelets or missing bags were recorded during these inspections. The i s o l a t i o n bags were l e f t i n place u n t i l the cones were picked. Cone c o l l e c t i o n Frequent checks of cone ripening state were begun i n late August to determine the optimum time of c o l l e c t i o n for each bag. Collection was made when browning of the scales of those cones v i s i b l e had reached the base of the v i s i b l e part, or when any of the cones had begun to open. Each bag was removed by cutting the branch below the label and carrying the whole intact to the laboratory. Whenever possible, a sample of open-pollinated cones was collected at the same time i n the v i c i n i t y of each bag, to permit a comparison of the seed y i e l d and the stage of development of the seed contents between the controlled and open-pollinated co l l e c t i o n s . The "check" cones were collected from the same whorl and on the same side of the tree, and usually from the same branch, as the bagged ones. Cones were removed, counted and placed i n labelled paper bags i n o o the laboratory, then placed i n cold storage (2 - 4 C) u n t i l commencement of seed extraction. A complete record of buds isolated and cones collected i s presented i n Appendix 10 . In some cases, eg. A x A, the number of cones harvested was greater than cone buds counted because of the slowness with which some buds swelled s u f f i c i e n t l y to permit certain i d e n t i f i c a t i o n . Where any doubt over the status or the health of a bud existed, i t was not counted during bagging. Cones were removed from storage and l e f t over night i n a warm room to hasten t h e i r opening. 204 Seed extraction Seeds were removed by hand. Each cone was checked c a r e f u l l y to ensure that a l l seed had been extracted. Wings were removed by hand rubbing and screening. The seed was passed through the forced-air column separator described previously, then separated into " f u l l " and "empty" seeds under X-ray fluorescence. Calculations of the percentage of f i l l e d seed and average number of f i l l e d seed per cone were based on counts of " f u l l " and "empty" seeds. Following t h i s , seed weights for each cross were determined after the rules of the International Seed-Testing Association (Voisenat, 1966): eight samples of 100 f i l l e d seeds were weighed after drying over calcium chloride for 36 hours. The number of samples and the number of seeds per sample had to be reduced for the selfed seeds because the f i l l e d - s e e d yields were so low. Seed X-raying In order to check on the completeness of the development of the seed contents, seeds from each cross and "check" c o l l e c t i o n s were X-rayed as described previously. Seed incubation Seeds were placed i n tap water i n small v i a l s stoppered with non-absorbent cotton and soaked for 36 hours at room temperature. They were placed on 8 cm. diameter paper germinator pads and placed immediately to incubate on a Jacobsen germinator apparatus manufactured by Zephyr N.V. Koel-en Luchttechniek, of Holland. Incubation conditions were: temperature 205 20° C constant, l i g h t 12 hours d a i l y from F40 Gro "Gro Lux" fluorescent tubes. Seed on each pad was spaced out approximately evenly to reduce the chance of fungal buildup and spread. No fungicide was used on the pads or i n the sub-irrigation water. The pads from each cone parent were arranged randomly i n one area of the germination table. Seeds from each parent were sown a few days apart to reduce the transplanting load. A l l seeds incubated had been X-rayed. A d a i l y check was made on germination. Seeds whose emerged ra d i c l e was equal to or i n excess of the seed coat length and whose r a d i c l e appeared normal and healthy were counted and removed d a i l y to separate "transplant" pads corresponding to each cross and r e p l i c a t e . Seeds pro-ducing abnormal r a d i c l e s , p a r t i c u l a r l y blunt, stunted or very f i n e , weak ones, were t a l l i e d separately, but not removed unless the r a d i c l e t i p evidently had died, or the tissue became infested with fungus, or u n t i l i t produced a root s u f f i c i e n t l y healthy to warrant t a l l y i n g as a transplant. Other abnormalities, such as the emergence of inverted embryos, were re-corded separately also. Any seeds badly invested with fungus were noted and removed. Germinator pads and wicks were changed when necessary i n an attempt to reduce the buildup and spread of fungi. The f i n a l count was taken at 50 days. Remaining seeds were cut open to determine whether any lacked an embryo. Germination per cent was based on the t a l l y of a l l seeds that had emerged contents (normal, stunted, p a r t l y germinated, inverted embryos) plus those with sound embryo but ungerminated at 50 days. Cumulative germination was plotted over time i n days, c l o s e - f i t t i n g , freehand curves were drawn for each cross and r e p l i c a t e , and the following values were obtained or derived for each graph: days to 206 culmination of rapid germination (germinative energy), days to 80 per-cent of f i n a l germination, and germination value (Czabator, 1962). These values were subjected to analysis of variance. Transplanting When s u f f i c i e n t seeds from each cross with length of emerged contents four times that of the seed coat had accumulated, seeds were transplanted with care into a 3 : 1 peat-vermiculite mixture i n styrofoam containers developed by the Canadian Forestry Service for the production of "plug-grown" stock. Each "plug" hole was 11.2 by 2.5 cm., tapered toward a narrower base, and comprised 2 cubic inches. Each f l a t contained 4 rows of 12 plugs. Each row was planted with 12 germinants of each cross, eg. A x A, A x D, A x E , A x wind, etc. Seedlings were chosen, as far as possible, to be equal i n size to those from the slowest-germinating source, i n an attempt to reduce biases possible i n favour of the earlier-germinating parents. Each seedling was dibbled i n to the base of the hypocotyl (marked by a colour change) and the medium pressed firmly around the root. Fine chipped granite g r i t was added to each plug to serve as a mulch. Wherever possible, f i v e replicates were established for each cross from each tree. F e r t i l i z e r was added i n solution on alternate days once the seedlings i n a tray had produced true leaves, whether or not the seed coat had r e -leased the cotyledons. The solution strength was adjusted to contain 100 ppm. of nitrogen. Otherwise, the trays were given a l i g h t d a i l y watering i n i t i a l l y using tap water, but a heavier watering less frequently as they grew t a l l e r . This was an attempt to avoid poor root growth i n a water-logged rooting medium. A continuous record of temperature and r e l a t i v e humidity was kept on a recording hygrothermograph i n the growth chamber. The position of each tray was randomly re-arranged weekly. Measurements Starting s i x weeks after transplanting, seedlings i n the growth chamber were observed for survival regularly and measured for height b i -monthly u n t i l apparently well after the culmination of r e l a t i v e growth rate of height. After 14 weeks the measurement i n t e r v a l was increased to 4, then 6 weeks. The f i n a l measurements were taken at 24 weeks. After f i n a l height measurements, 4 seedlings nearest i n height to the mean of the 6 t a l l e s t from each re p l i c a t e of each cross were removed for detailed study. The following counts and measurements were obtained: epicotyl length to base of l a s t leaves, hypocotyl diameter at the coty-ledons, number of branches, length of longest branch and t o t a l branch length. The seedlings then were cut at the cotyledons and oven-dried at o 105 +2 C for 48 hours. Leaves, branches and stem were separated and dry weights of each, as well as roots (including hypocotyl), were measured after reaching room temperature over CaCl 2 i n desiccators. Linear re-gressions and analysis of variance were conducted on these parameters, on combinations of them, or r a t i o s calculated from them. Seedlings removed from the o r i g i n a l germinator pad but not trans-planted to the growth chamber were allowed to develop to the state of free cotyledons on the transplant germinator pads on the incubating unit. 208 Once i n that stage they were measured or assessed for the following features: cotyledon, number, length of a single cotyledon, hypocotyl length and hyopcotyl colour. Colour was entered as one of the following classes: Class No. Colour Class No. Colour Yellow Yellow-green 6 Brown Red-Brown 3 Green 8 Red Green-brown Deep Red Light brown The frequency by classes was calculated for each cross. Regression analyses were conducted on the cotyledon and hypocotyl length measurements. 209 4.22 Results and discussion Receptive date Tree A was receptive f i r s t and tree E l a s t . The approximate dates of optimum r e c e p t i v i t y were: tree A, A p r i l 2, tree D, A p r i l 3, tree E, A p r i l 6. More detailed observations, and comments on strobilus appearance are pre-sented i n Appendix 11 Cone maturity and harvest Cones matured f i r s t on tree A and l a s t on tree E, although there was considerable overlap. Collections were made from tree A between September 22nd and October 12, with most bags harvested by September 29th. Tree D was harvested mostly on September 24th; only one bag remaining u n t i l October 2nd. Tree E coll e c t i o n s began on September 29th and lasted u n t i l October 14th, with most bags collected on the e a r l i e s t date. The mean time from r e c e p t i v i t y to cone maturity i n the bags, by parent tree, were: A, 181 days, D, 175 days, E, 177 days. This indicates that they responded to three d i f f e r e n t environments, or that they are s l i g h t l y d i f f e r e n t genetically and responded d i f f e r e n t l y i n similar environments. Bag y i e l d The y i e l d of bags and cones i s presented i n Appendix 10 . Loss of bags was not serious, only f i v e from tree A and one each from the others out of a t o t a l of 62 bags. Those l o s t were mainly on twigs apparently too fine to withstand the bag motion i n high winds. 210 Cone y i e l d Cone abortion was very s l i g h t . Where abortion occurred, no pattern associated with p o l l e n parent was apparent. Damage to cones, branches and bags by birds and insects was l i g h t . Only a few earwigs were found i n one bag on tree D. The l i g h t bag l o s s and cone abortion found here compare favourably with those known f o r western hemlock (Piesch, 1974), and those reported f o r other c o n i f e r s (egs. Bingham and S q u i l l a c e , 1955, Barnes et a l . , 1962, Orr-Ewing, 1954, S z i k l a i , 1964). These r e s u l t s are based on only three trees and one year's study, so they may be a l t e r e d for other trees and other years. The weather during the summer included high winds on a few occasions, no doubt exerting considerable s t r a i n on the bagged branches, yet they withstood i t quite w e l l . No published information was found on the natural abortion rate of western hemlock conelets. A cursory examination of the branches near the bags at cone harvest indicated that i t was not high during 1970, but many conelets could have been l o s t before they became woody. I s o l a t i o n e f f i c a c y Two hundred and t h i r t e e n non-pollinated " c o n t r o l " cones were c o l l e c t e d , y i e l d i n g 5530 fully-developed (rounded) seed. Only eight f i l l e d seeds were found during fluoroscopy. A l l came from a s i n g l e bag on tree E. These were X-rayed separately and appeared normal. Open-pollinated cones c o l l e c t e d near the " c o n t r o l " bags a l l yielded f i l l e d seeds. 211 The bag containing the f i l l e d seeds was isolated on March 20th, the same day on which pollen shed from one strobilus on an adjacent tree (A) was discovered. Sixty-three cones i n four other control bags on tree E established the same day contained no f i l l e d seeds. The contamination by foreign pollen appears to be s l i g h t , and confined to tree E. Seed y i e l d The gross yields of f i l l e d seeds and i n percentage and per cone, are presented by cone and pollen parent i n Appendix 13 , and i n summarized form i n Appendix 14 . The results were analysed by ANOVA and tested further by Duncan's t e s t . These groupings are indicated i n Appendix 14 Self p o l l i n a t i o n reduced f i l l e d - s e e d y i e l d strongly i n each case (Appendix 14 ). Wind p o l l i n a t i o n produced yields intermediate to s e l f and outcross dry p o l l i n a t i o n , which gave the highest y i e l d s . Wet p o l l i n a -t i o n reduced seed y i e l d from outcrossing i n tree E to half that from dry outcrossing, but to the l e v e l of dry se l f i n g i n tree A. A l l data of controlled outcrossing, s e l f i n g and wind p o l l i n a t i o n were combined and tested against each other. A l l differences were s i g n i f i c a n t . No difference i n seed y i e l d occurred between cone or pollen parents when only controlled, dry outcrossing was considered. The yiel d s from s e l f p o l l i n a t i o n s did not d i f f e r . Yields per cone from reciprocal outcrosses were not di f f e r e n t . Comparisons of the seed y i e l d from selfings pollinated by hand or i n shaker bags were made to compare the p o l l i n a t i o n efficacy of the two methods (Table 4-4 ). 212 Table 4- 4 Self-pollinated seed y i e l d by p o l l i n a t i o n method. Trees A and E. Hand p o l l i n a t i o n Shaker p o l l i n a t i o n Tree Cones S e e d s % F u l l Cones S e e d s % F u l l Total F u l l F u l l /cone Total F u l l F u l l /cone A 18 598 50 8.4 2.8 16 489 59 12.0 3.7 20 553 78 14.1 3.9 16 450 49 10.9 3.1 Mean 11.1 3.4 11.5 3.4 E 24 374 22 5.9 0.9 32 939 108 11.5 3.4 18 330 11 3.3 0.6 20 315 19 6.0 0.9 28 456 14 3.1 0.5 20 293 11 3.8 0.5 Mean 4.0 0.7 5.5 0.8 Differences between the means were too small to warrant s t a t i s t i -c a l t e s t s , p a r t i c u l a r l y since v a r i a t i o n within samples was greater than between means. No apparent difference exists between the effic a c y of the two methods i n producing f i l l e d seed. Therefore, yie l d s of f i l l e d seeds from "shaker" self bags do not seem to be limited by a poor d i s t r i b u t i o n of pollen or dichogamy. Open-pollinated yi e l d s d i f f e r e d and were not consistent. Using f i l l e d seed per cone, A was superior to D and E, but using percentage of f i l l e d seed, E was equal to A. This i s because E produced a lower number of "round" seeds ( f u l l - s i z e d , p o t e n t i a l l y f i l l e d seeds) than the others. From each cone, many flattened and smaller seeds were recovered. The small and large seeds often occurred together on each scale. This was not so i n the other trees where small, s h r i v e l l e d seeds were found only toward the base of the cones. The low open-pollinated y i e l d from tree D might be a result of i t s relative, i s o l a t i o n . It.stands on the edge of a grove of trees, open to a l l but the south-west quarter. But i t i s about 500 feet from the nearest tree to the northwest, from which d i r e c t i o n drying winds are most l i k e l y to blow during p o l l i n a t i o n time. Dry p o l l i n a t i o n generally gave much better seed yi e l d s than wet p o l l i n a t i o n , although tree E did not exhibit the same depression i n y i e l d as tree A. This reduction i n y i e l d might be due to a lower concentration of pollen near the micropyle, resulting i n poorer "germination" of the pollen, as suggested by McWilliam (1960), or i t might be the re s u l t of irregu l a r pollen development produced by the d i s t i l l e d water used. Among the many exogenous influences affecting germination and development i n v i t r o are temperature, pH +, oxygen, osmotic pressure, carbohydrates, and various anions and cations (Stanley, 1967). Two ions that seem p a r t i c u l a r l y important for proper germination and development are boron and calcium. Ho and S z i k l a i (1971) obtained germination and development to the male game-tophyte stage of western hemlock i n a solution containing both boron and calcium ions i n d o u b l e - d i s t i l l e d water. D i s t i l l e d water was used i n these wet p o l l i n a t i o n s , but i t was not mixed with the pollen u n t i l just before wet p o l l i n a t i o n s were to be done on each tree so that the pollen could not imbibe much before being released. Perhaps s u f f i c i e n t imbibition occurred to stimulate pollen development too early. Stanlake and Owens (1974) found that nucellar growth resulted i n contact with pollen grains along the microylar canal. Perhaps the solution caused the micropyle to close quickly, excluding most pollen grains, or water surface tension may have 214 prevented the pollen grains penetrating far enough to be ef f e c t i v e . Cone size The observations i n Appendix 12 indicate that the bagging caused noticeable differences i n cone size to develop quickly, reaching perhaps two to four times the volume of nearby unbagged cones, e.g., trees A and D on A p r i l 28th. In most cases the differences were much less noticeable at cone harvest. The results from the measurements taken on bagged and unbagged cones appeared i n Chapter 3, Section 3.1. Seeds of two cone collections from tree A were extracted by indi v i d u a l cones to permit a comparison of the number of round, and p o t e n t i a l l y f i l l e d , seeds per cone from bags with those open to the a i r . The means and standard deviations are: bagged 28.65 seeds + 6.62; unbagged 25.64 seeds + 5.40. The difference i s not s i g n i f i c a n t . A record of f u l l , empty round, and f l a t seeds was kept by cone for A x A hand-pollinated cross. (Table 4-5 ). Table 4- 5 Y i e l d of f u l l , empty and aborted seeds following s e l f i n g of Tree A Round seeds/ cone - f u l l 2 5 4 2 5 3 6 6 4 4 6 2 3 1 5 3 4 3 2 3 - empty 16 25 22 29 25 25 26 23 19 34 18 18 18 18 17 15 18 25 17 17 F l a t seeds 3 4 1 0 2 4 5 2 2 0 1 3 1 3 2 3 2 6 3 2 Totalround 18 30 26 31 30 28 32 29 23 38 24 20 21 19 22 18 22 28 19 20 Total poss-i b l e seeds 21 34 27 31 32 32 37 31 25 38 25 23 22 22 24 21 24 34 22 22 215 The average number of round seeds per cone was 24.9 + 5.61, and of t o t a l possible 27.3 + 5.68. The fact that every cone produced at least one f i l l e d seed indicates that pollen coverage by hand spraying was f a i r l y good. A plot of the number of f i l l e d seeds per cone over t o t a l seeds indicated no strong tendency for the bigger cones to produce more selfed seeds. Similar results were obtained for a wet cross bag (E x A), for which the mean y i e l d was 15.7 f i l l e d seeds per cone. In another wet cross ( E x D), each cone yielded some f i l l e d seed, indicating again that the coverage of the s t r o b i l i during p o l l i n a t i o n was adeguate. Number of scales per cone Samples of bagged and unbagged cones from each tree were examined for number of scales to determine the maximum number of seeds possible and provide a basis of comparison for actual and potential seed yi e l d s per tree. The re s u l t s appear i n Table 4-6 . Table 4-6 Mean number of scales for bagged and unbagged cones B a g g e d U n b a g g e d Tree Mean S.D. C.V.% Mean S.D. C.V.% Bagged-; unbagged A 24.3 2.82 11.6% 23.2 1. 94 8. 3% 1. 05 D 26.1 2.34 8.9% 24.6 2. 79 11. 3% 1. 06 E 28.1 3.69 13.1% 24.4 1. 76 7. 2% 1. 15 I t appears that bagging i s associated with an increased number of scales, but i n only one case (tree E) was the difference s i g n i f i c a n t . Probably the apparent difference i s due to the fact that the bags were situated where the s t r o b i l i were most dense and near the t i p s of l a t e r a l branches. In Chapter 3 i t was found that cones i n a terminal position or 216 toward the t i p s of main limbs were larger than cones fart h e r back, which us u a l l y provided the open-pollinated c o l l e c t i o n . The p o t e n t i a l number of seeds per cone (twice the number of scales) i s 48 f o r tree A, 52 for tree D, and 56 f o r E. These are well above the numbers of f i l l e d seeds per cone derived from dry c r o s s - p o l l i n a t i o n s : 21.5 for tree A, 23.4 f o r D, and 18.7 for E, which are, r e s p e c t i v e l y , 76.4, 86.2 and 82.3 percents of the round seeds per cone. Dissection of a number of bagged cones indicated that the number of " s t e r i l e " basal scales ranged from 6 to 11, averaging 7.8 + 0.84. This reduced the median number of seed-bearing scales to 18.2, and the p o t e n t i a l seed crop to approximately 36 per cone. The y i e l d of f i l l e d seeds per cone ranges between 51% and 64% of that p o s s i b l e . Thus there i s some p o s s i b i l i t y of improving the seedling y i e l d from western hemlock cones, and study of the v a r i a t i o n i n f i l l e d seeds per cone from c o n t r o l l e d crossing and the possible f a c t o r s leading to scale and seed s t e r i l i t y might be f r u i t f u l before lar g e - s c a l e plans f o r c o n t r o l l e d breeding are made. Fowells (1965) gives t o t a l seed y i e l d f o r western hemlock as "...approximately 30 to 40 seeds..." per cone. These trees f a l l within h i s l i m i t s , based on number of " f r u i t f u l " scales, but below i t when " t o t a l round" seeds are considered: 28 f o r tree A, 27 for D and 23 f o r E. With only three trees examined, i t i s possible that chance alone could account fo r t h e i r lower f i g u r e s and that the population mean f a l l s within the l i m i t s given by Fowells. The only data on the y i e l d of f i l l e d seeds per cone following con-t r o l l e d p o l l i n a t i o n of western hemlock were obtained from Langner (1972) 217 and Piesch (1974). From two trees of western hemlock, Langner's range of f i l l e d seeds per cone from outcrosses was 5.0 to 9.3, averaging 6.8. Self-pollinated cones yielded from 1.7 to 3.3 f i l l e d seeds, averaging 2.7. No f i l l e d seeds were found among those collected from unpollinated "control" bags. Piesch (1974) reported y i e l d s of f i l l e d seeds per cone between 12.1 and 20.8, averaging 16.1, from outcrosses, and 0.4 to 3.6, with a mean of 1.5, from s e l f p o l l i n a t i o n s . Self p o l l i n a t i o n s by "shaker" bags were s l i g h t l y ahead of s e l f s produced by syringe p o l l i n a t i o n : 1.6 vs 1.2 f i l l e d seeds per cone. Unpollinated "control" bags yielded more f i l l e d seeds per cone than s e l f s , indicating that the results are not wholly r e l i a b l e . Those data show more v a r i a b i l i t y and lower f i l l e d seed y i e l d from controlled outcrossing than do these from Totem Park. Piesch used similar methods to those employed here, except that paper i s o l a t i o n bags and absorbent cotton were used. No information on the amount of dry pollen applied per bag, the stage of r e c e p t i v i t y of the s t r o b i l i , or of the kind of bag used or the age of the pollen were provided by Dr. Langner. But i t i s most l i k e l y that the work was done by experienced tree breeders, so that these potential sources of poor seed y i e l d were reduced to the minimum. Perhaps the low y i e l d from Langner's crosses stems from a narrow genetic base of the population from which the trees were drawn. If the parents used here are half sibs, f i l l e d - s e e d y i e l d may have been reduced somewhat by the greater degree of inbreeding produced. In view of Langner's (1972) r e s u l t s , and the confinement of f i l l e d seeds i n the "control" bags here to only one bag, i t appears that western hemlock i s not apomictic. However, i t may occur on certain trees, or on these trees i n other years (Orr-Ewing, 1957b); establishment of such 218 "control" bags i n further breeding t r i a l s i s recommended. Great v a r i a b i l i t y i n filledr-seed y i e l d from breeding t r i a l s can occur from year to year (Snyder and Squillace, 1966, Hyun, 1969), so the same might be found for western hemlock. Perhaps the heavy pollen a p p l i -cations i n t h i s t e s t boosted the seed y i e l d inordinately. But Hyun (1969) demonstrated that f i l l e d - s e e d y i e l d from a hybrid cross (Pinus r i g i d a x taeda) did not change s i g n i f i c a n t l y when pollen applied per bag was reduced from 1.0 to 0.3 cc. I f the same i s true for western hemlock, considerable economy of p o l l i n a t i o n time or of valuable pollen, through d i l u t i o n with dead pollen or other materials, could be real i z e d . The trees used i n t h i s study were large, open-grown and full-crowned, and produced a bumper cone crop i n 1970. The following year they were almost barren. Consequently, the number of cones per bag probably i s higher than that obtainable over a period of years. Also, the y i e l d of f i l l e d seeds per cone l i k e l y w i l l vary with year and location. However, the high bag survival and low cone abortion, combined with the high number of cones per bag and f i l l e d seeds per cone found here and by Piesch (1974) , indicate that one or two bags w i l l be enough to test a cross i n both the laboratory and the f i e l d . Seed weight This feature was studied because of i t s potential influence on early growth. The mean weights of 100 seeds by cross are presented i n Table 4-7. Although mean weight from each tree d i f f e r e d s i g n i f i c a n t l y , ANOVA indicated that a s i g n i f i c a n t interaction between cone and pollen parents occurred, making i t improper to conclude anything about the ov e r a l l means. 219 Table 4-7 Mean fresh * weight per 100 seeds (grams) by cone and p o l l e n parent. P o l l e n P a r e n t Cone Parent A D E Wind .2432a .2237b .2500ab .2458bc .2365b .2331b Cone Mean S.D. C.V.% .2301a .0054 2.35 .2485b .0062 2.49 .2570c .0083 3.23 A .2378a .2184c D .2429c .2554a E .2729a .2757a Pollen Mean .2510a .2474ab .2442ab .2342b .2445 * Weight a f t e r drying 36 hours over CaCl^ at 34 + 2°F and warmed to room temperature. ** Means within any cone parent sharing the same l e t t e r not s i g n i f i c a n t l y d i f f e r e n t at the 5% l e v e l . *** Mean value from two bags that d i f f e r e d s i g n i f i c a n t l y . Further comments below. Inspection of the table reveals the reason: no p o l l e n parent i s co n s i s t e n t l y associated with the heaviest or l i g h t e s t seeds, although wind-p o l l i n a t e d seeds u s u a l l y are l i g h t e r than the others (0.2342 gm. vs. 0.2473 gm.). This was tested against a l l c o n t r o l - p o l l i n a t e d seed and found s i g n i f i c a n t . V a r i a b i l i t y i n seed weight within the tree i s small i n every case, but d i f f e r e d s l i g h t l y between trees, increasing with seed weight. The reason that greater v a r i a b i l i t y was found i n tree E may be li n k e d to the high number of bags established: 27. This required more space and, conse-quently, r e l a t i v e l y more of the crown i n t h i s tree. Crosses A x D and A x E were represented at harvest by only two bags 220 each. Each of the A x D bags yielded enough f i l l e d seedsto permit a check on the possible v a r i a b i l i t y between bags i n seed weight and germinative behaviour. The mean weight from the 22-cone bag was 0.2146 gm. per 100 seeds, that from the 28-cone bag was 0.2222 gm. The difference, although small, i s s i g n i f i c a n t . Selfing tree E produced l i g h t e r seeds than crossing i t with A and D. The selfed weight i s based on 10 i s o l a t i o n bags distr i b u t e d throughout the crown, which should reduce the non-random influence that portion of the crown might exert on seed weight. Therefore, there i s a p o s s i b i l i t y of xenia i n seed weights. The other s e l f s produced heavy seeds,and one cross, D x A, gave l i g h t seeds, so xenia, i f i t e x i s t s , may d i f f e r with each cross. Discussion, seed weight The fact that a f i v e percent difference i n seed weight between open and controlled p o l l i n a t i o n s exists must be taken into account when measure-ments of early growth are compared. Like l y the reason for the difference i s the protection afforded by the i s o l a t i o n bags during the f u l l growing period. The strongest influence on seed weight i s cone parent, doubtless because i t produces the gametophytic tissue ("endosperm") around the embryo and t h i s tissue constitutes the bulk of the seed weight (Allen, 1961). Of the pollen parents only tree A gave greater mean weights from controlled than from open p o l l i n a t i o n , although t h i s was due mainly to tree E. This weak evidence for xenia i s s i m i l a r to that found for seed weight i n western white pine by Squillace (1957) and the a r t i c l e s reviewed e a r l i e r , A comprehen-sive test w i l l have to be established and analysed before i t can be ruled out of western hemlock studies. But i n the meantime, d i s t r i b u t i n g bags for 221 any combination around the crown and bulking the seeds w i l l protect against non-randomness i n the maternal influence on seed weight. Seed germination Total germination The values of germination percent at 50 days are presented i n Table 4-8. Table 4-8 Germination percentage at 50 days by cone and pollen parent Cone parent A P o l i e n D P a r E e n t Wind Total f i l l e d seeds Weighted Mean** A D 83.3 82.7 (22) (28) *85.3 *84.3 78.0 89.8 93.4 85.7 90.7 1274 1129 86.57 88.66 E 72.0 79.7 69.9 76.8 1271 75.53 Total Seed 912 776 886 1100 3674 Weighted mean 78 .61 81.70 88.60 84.36 83.39 * "22" and "28" refer to seeds maintained separately from 22-cone and 28-cone crossing-bags. ** Weighted by the number of f i l l e d seeds i n the t e s t . The data were tested for s i g n i f i c a n t differences due to cone and pollen parent. Tree E was lower than the others, reaching only 75.5% vs. 86.6% for tree A and 88.7% for tree D. Neither pollen parent nor cone x pollen parent interaction had a s i g n i f i c a n t influence on t o t a l germination. 222 Tree D selfed germinated less than D x E and D open-pollinated, but the low number of selfed seeds available (50) weakens the comparison. No difference was found between the bags producing A x D seed. Germinative rate a) Germinative energy This ..feature was, analysed for evidence of genetic control, since i t was used i n the population study (Chapter 3). Calculation of t h i s value i s i l l u s t r a t e d i n Figure 4-6 . Mean values are presented i n Table 4-9 . Table 4- 9 Mean germinative energy by cone and pollen parent (days) Cone P o l l e n P a r e n t Weighted* parent A D E Wind mean 1 A 25.0c (22)** 21.8b 18.2a 20.0ab 20.54 b (28) 18-.0a D 18.7b 14.7a 18.5b 18.7b 17.87 a D 32.7b 31.2b 23.8a 30.0b 29.12 c Weighted* mean 25.50b 21.65a 20.46a 22.92ab 22.50 1 Common l e t t e r s within cone parents indicate s t a t i s t i c a l l y similar groupings (P= 0.05). * Weighted by the number of "normal" seed counted. ** Values from bags containing 22 and 28 cones, respectively. Tree E obviously i s the slowest, but a s i g n i f i c a n t interaction occurred between cone and pollen parents i n the analysis of variance, so no conclusion should be drawn about the pooled means. The selfed seeds behaved e r r a t i c a l l y - those from tree A were slower than the mean of t h e i r h a l f -sibs, while those from trees D and E were faster. Reciprocal crosses i n -223 100 w CD <D cn c o •H +J ca c •H e U cu cn CD > •H 4-J CO rH e u 90 80 r-70 60 50 40 30 20 10 v-- 0 Figure 4-6. Reprsentative curves of cumulative germination of E x D seeds, showing calculation of germinative energy. 224 volving A and D behaved consistently, while any involving E did not. Wind-pollinated seeds were always close to the mean. Mean values by replicate ranged between 14 and 47 days, the l a s t from E x A r e p l i c a t e 4, where no d e f i n i t e break i n the curve occurred. A s i g n i f i c a n t difference occurred between the values from the 22-cone and the 28-cone bags containing the A x D cross. b) Days to reach 80% of f i n a l germination Table 4 - i o presents these data. Cone parent was the only factor s i g n i f i c a n t l y affecting t h i s estimation of germinative rate. Each tree di f f e r e d s i g n i f i c a n t l y from the others, with D the fastest and E the slowest. Tree D was also the most consistent i n i t s rate, judging by i t s low c o e f f i -cient of v a r i a t i o n . Although D selfed germinated less than the D crosses, the seedlings did not emerge slower. The other s e l f s did not diverge strongly from the general mean for t h e i r trees, although, i n each case, the rate appeared somewhat lower than those of the other crosses. No difference was found between the bags producing A x D seed. Table 4 - i o Days to reach 80% of'-final germination.;by cone and pollen parent. P o l l e n P a r e n t Cone Weighted* Co-eff. of Parent A D E Wind mean Variation % A 21.3. (22)**..21.4 18.3 18.9 19.25b 16.61 (28) 18.6 D 13.3 14.0 13.6 13.6 13.56a 4.41 E 32.3 30.7 34.8 29.9 31.37c 17.33 Weighted mean 23.05 - 24.83 18.57 20.88 21.69 * Weighted by number of "normal" seeds counted. ** Values from bags containing 22 and 28 cones, respectively. 225 c) Germination Value The value proposed by Czabator (1962) was calculated for each cross and replicate. Values obtained are presented in Table 4-11. Table 4 r - l l Germination Value for cone and pollen parent Cone P o l l e n P a r e n t Parent A D E Wind M e a n A D E 2.73de 2.34e 7.54abc 6.95bc 4.19b 6.43c 8.49ab 8.76a 7.35abc 7.71a 3.76d 4.23d 3.80d 4.19d 3.99b Mean 4.30 b — 3.87 — b — 6.48 — a 6.16 — a 5.10 The letters accompanying the values throughout the table indicate similar results as determined by Duncan's test. The direction of comparison under .the "Mean" headings is indicated by the lines. They indicate the super-i o r i t y of Tree D as a cone parent. However, ANOVA demonstrated that a significant cone x pollen parent interaction occurred, so i t is not possible to draw safe conclusions from main comparisons. The interaction i s caused mainly by the low value of the A x D cross versus the higher A x E and A x Wind crosses and the high value of D selfed. Differences existed between reciprocal crosses, for example A x D versus D x A, and A x E versus E x A, Open pollination gave values near or above the mean for each tree. The high value for D selfed was due mainly to i t s rapid germination, rather than higher germination. Comparing the two bags producing A x D seeds,the 22-cone bag was significantly poorer than i t s 28-cone counterpart (1.60 vs. 226 3.08), mainly because of slower emergence. No correlation was found between germination percent or rate and 100-seed weight. Since t o t a l germination percent at 50 days was not affected by s e l f i n g , i t seems that f i l l e d seeds from s e l f i n g are as capable as outcross seed of developing to the stage considered here as "germinated". Piesch (1974) obtained similar r e s u l t s . That suggests that the gene systems involved i n mobilizing reserves and elongating the embryo to reach "germin-ation" are not under such strong elimination pressure as those involved i n producing a viable embryo-female gametophyte complex ( " f i l l e d " seed). Thus, western hemlock seems to d i f f e r from slash pine (Squillace and Kraus, 1962), Picea abies, and Larix decidua, and Pseudotsuga menziesii (Franklin, 1970). The d i f f i c u l t y of estimating germination rate i s demonstrated by the inconsistency of results from the three measures used: cone x pollen parents gave a s i g n i f i c a n t interaction i n two of the cases. (Although Czabator's germination value combines speed and completeness of germination, i t i s discussed here). Germinative energy and Czabator's method employ the same point on the Curve: that where the trend of cumulative germination deflects away from the v e r t i c a l . I f such a point does not occur during the test period, the method i s not appropriate. The use of the "days to reach 80 percent of f i n a l germination,