Relation of Crown and Foliage Traits to Height Growth of Sitka Spruce D.L. Copes, USDA Forest Service, Pacific Northwest Research Station, Forestry Sciences Laboratory, 3200 SW Jefferson Way, Corvallis, OR 97331; W.H. Pawuk, USDA Forest Service, Tongass-Stikine Area, Box 309, Petersburg, AK 99833; W.A. Farr (deceased), USDA Forest Service, Pacific Northwest Research Station, Forestry Sciences Laboratory, 2770 Sherwood Lane Suite 2A, Juneau, AK 99801-8545; and R.R. Sileo (retired) USDA Forest Service, Pacific Northwest Research Station, Forestry Sciences Laboratory, 3200 SW Jefferson Way, Corvallis, OR 97331. ABSTRACT. Four crown and foliage traits of a young Sitka spruce (Picea sitchensis) stand were tested with rooted cuttings in greenhouse and field plots for possible thinning selection guidelines. Repeatability estimates of the amount of genetic control over the four traits and the relation of those traits to height growth were evaluated 5 years after rooting. Only the blue-green trait was significantly associated (P < 0.05) with height growth. Average height of cuttings selected for green foliage was 17% greater than cuttings selected for blue foliage. Under greenhouse conditions, repeatability estimates of green or blue foliage were 98%versus 66%, respectively; 72% versus 89%for trees with dense or open crowns, respectively; 65%for the comparison of upright and horizontal branches; and 75%for the comparison of long or short branches. West. J. Appl. For. 11(3):77--80. C oastal Sitka spruce (Picea sitchensis) forests in southeast Alaska develop natural regeneration soon after harvest (Har­ ns and Farr 1974). Stocking typically averages 4000 stems per acre, often with patches of 20,000 stems per acre. Nearly all clearcut areas need precommercial thinning to reduce stocking when the trees become 4 to 6 m tall. Because of this abundance of natural regeneration, it may be possible to use thmning to achieve a small amount of genetic improvement m Sitka spruce populations if growth rates are linked to easily observable traits used in thinning guidelines. Thinning programs in southeast Alaska currently use a combination of spacing control, species selection, and phe­ notypic selection. Because of their higher value, Sitka spruce, and Alaska-cedar (Chamaecyparis nootkatensis), and west­ em red cedar (Thuja plicata) are strongly favored over western hemlock (Tsuga heterophyla). Traits with strong genetic control are more useful as selection traits. We know that heritability estimates of height growth range from 0.52 to 0 82 (family) and from 0.11 to 0.34 (individual) (Faulkner 1987). These results indicate moderate to good genetic con­ trol and, based on heights of 6 yr old trees, project l 0 to 22% gams. Thus, all thinning favors taller specimens. To maximize genetic change or gain by selection, all easily differentiated and highly variable traits with high heritabilities must be identified. Crown shape and size in Sitka spruce result from variation in traits such as branch length, branch angle, number of branches per whorl, and number of internodal branches (Cannell 1974 and 1982), but little is known of the degree of genetic control of branch, foliage, or crown traits and any relatedness of crown traits to growth traits. This paper reports the results of using rooted cuttings to test the genetic control and relation to height growth of the only four highly contrasting crown and foliage traits that appeared practically observable in thinning-aged stands. Study Areas and Methods Trait Selection In 1985, several hundred 22 to 25 yr old Sitka spruce trees growing in a well-stocked, naturally regenerated stand of spruce and hemlock in southeast Alaska were scrutinized for potential thinning traits. The area surveyed was about 17 k south of Petersburg on Mitkof Island. We sought to identify easily WJAF 11(3) 1996 77 discerned, contrastmg crown and foliage trmts that could be used as criteria in precommercial thinning guidelines. Observations of crown traits revealed only four that were both easily identified visually and highly variable in the population. The traits were: dark bluish-green (called blue in this paper) or normal green foliage, wide or narrow crown (long or short branches), upright or horizontal branch orientation (branch angle), and dense or open crown density. Trees with wide crowns had branches that were 30 to 50% longer than similar branches on trees selected for narrow crowns. Trees selected for dense crowns were trees that an observer could not see through the canopy, while an observer could easily see through the canopy of trees selected for open crowns. Tree selection was based solely upon visual observa­ tion. Seven pairs of trees with contrasting phenotypes (visible characteristics) were selected for each of the four traits. To facilitate visual comparison, trees of each pair were neighbors. Propagation and Assignment of Test Trees Branch tips were collected in January 1986 from the 56 selected trees and rooted at Corvallis, Oregon, and at Peters­ burg, Alaska. Rooting followed the procedures outline by Copes (1987). In November 1986, the rooted cuttings at Corvallis were lifted bareroot and transferred to the green­ house at the Petersburg nursery, where the rooted cuttings from both locations were transplanted into 2 gal containers filled with potting mix. The cuttings were grown in the greenhouse during the 1987 growing season under standard nursery procedures for conifers. In September 1987, the cuttings were sorted into two groups for the field test and the greenhouse test. Enough cuttings were available to evaluate three traits with six pairs of trees per trait in both environments. Cuttings observed for the branch length trait were evaluated with only five pairs of trees. Twenty cuttings per tree were allocated to the field, and 17 to 20 cuttings per tree were used in the greenhouse. Greenhouse Test The greenhouse was divided into two blocks, and each block was divided into four trait subplots. For each paired comparison, 8 to 10 rooted cuttings of each tree were placed in the appropriate trait-subplot of each block. Several fertil­ izer applications were made during the 1989 and 1990 grow­ ing seasons, but none were made in 1991. The 5 yr old cuttings were evaluated in September 1990. Tree height was measured to the closest 0.5 em. Foliage color and crown density were visually scored as blue or green and dense or open, respectively. Early assessment of repeatability of branch length (crown width) and branch angle required measurements because relative differences in crown struc­ ture were difficult to see in small trees. The mean length of the longest three branches in the 1990 whorl of each pair of cuttings was determined, and a qualitative assessment of trait repeatability was calculated from the branch measurements. Repeatability in this case was the percentage of paired cuttings of the trees that exhibited the same relative rank­ order difference (larger or smaller) as their parent trees. In the same manner, repeatability of upright or horizontal 78 WJAF 11(3) 1996 branch angle was calculated from measurements of the average angle of the three longest branches in the 1990 whorl (measured to± 5 '). Field Test Rooted cuttings for the long-term test were planted m the field in early June 1988 on a clearcut unit at the same elevation located 3 km south of the parent trees. The areq was divided into 2 block,s and each block was divided into 4, randomly assigned trait-subplots. In each block, 20 cuttings of each pair of parent trees (10 per tree) were randomly planted in adjacent rows at 3 by 3m spacing. The 7 yr old cuttings were measured in September 1992 after 5 yr of growth in the field. Measurement was limited to cuttings of trees selected for blue or green foliage because the greenhouse test revealed that only trees with that trait exhibited a significant relation with height growth. The same visual assessment of blue or green made in the greenhouse test was done in the field. In addition, tree height and annual leader increments from 1990 through 1992 were recorded for each cutting. All cuttings that were severely suppressed by competing vegetation or exhibited atypically slow growth due to extremely poor microsites were excluded from measurement. Statistical Analysis Data from the greenhouse and field tests were subjected to analysis of variance (ANOVA) using the SAS proce­ dures for general linear models (SAS Institute Inc. 1989) The experimental design was a split plot (mixed model) Contrasting pairs and the traits were random and fixed effects, respectively. Separate analyses were performed for each trait. The independent variables investigated m the greenhouse test were the four crown or foliage traits, but based upon greenhouse results, only the green or blue trait was measured and analyzed in 1992 for the field test Transformation of data before analysis was not necessary Differences were judged significant when the probability of obtaining a larger F-value by chance was ::; 0.05 Sources of variation and expected mean squares of the ANOV As were: Source D F Blocks (B) B-1 Pairs (P) P-1 B (B-1)(P- 1) x P e( rror 1) Phenotypic trait (H) P xH e( rror 2) Residual (error 3) H -I Expected mean squares (j + h p + ph (j + h p + bh; (j + hp 2 (JE 2 h 2 l:8 +b-<JPH +bp b-1 b-1 h -- 2 (P-l)(H - I) cre +bHcrPH P(B-l )(H - I) cr2 E Error 2 was used to determine the F-value for phenotype (H) because the individual pairs of contrasting cuttings were the experimental units of treatment. The total degrees of freedom were 23 and 19 when the tests had 6 and 5 pairs of trees per trmt, respectively. Results Discussion Trait Repeatability The relation between foliage color and leader growth was significant in both the greenhouse and field environments. If an increase in average height growth is desired and species, spacing, and other selection criteria reveal no advantage of one tree over another, field workers should select Sitka spruce trees with green foliage. The actual gain in height growth obtained from such selection will be less than the 17% obtained in this study because only a fraction of the opera­ tional thinning choices will permit blue or green selection. We selected only the trees with the darkest blue foliage for comparison with trees with green foliage so the growth potential of trees with intermediate foliage color is not known. A known physiological explanation exists that partially explains why conifers with green foliage color may grow taller than those with blue foliage. Clark and Lester (1975) report that relative photosynthetic capacity of Colorado blue spruce (Picea pungens Engelm.) decreases when blue foliage increases. Salisbury and Ross (1978) show that blue-green foliage reflects the higher energy blue light wavelengths (400 to 500 nm), whereas trees with green foliage absorb more than 90% of those wavelengths. When other factors control­ ling growth are equal, trees with blue foliage will absorb less light energy and may not support as much photosynthesis and growth as trees with green foliage. This simple explanation knowingly ignores many other variables that contribute to growth and other important traits. The other crown traits were not significantly associated with height growth at 5 and 7 yr, but we feel that it is too early to eliminate them from future study. The annual accumula­ tion of small differences over several decades may result in large differences in photosynthetic area. The 21% difference per year in average growth of branches (5.2 em) was not significant at age 5, but that same difference accumulated annually for 20 or 30 yr may result in significant differences and trees with large crowns may produce more diameter growth. A similar change over time may occur with the other crown traits. The amount of genetic control (repeatability) over the four traits differed greatly. Repeatability of green fohage was high in both environments, while repeat­ abtlity of blue foliage was high in the field but low in the greenhouse due to irregular trait development in cuttings of two blue trees grown under low nitrogen conditions. The percentage of cuttings that exhibited blue or green foliage identical of their parent was 90% and 94% in the field, respectively, and 66% and 98% in the greenhouse, respectively (Table 1). Repeatability estimates for cuttings from dense or narrow crown trees were 89 and 72% respectively (Table 1) Cuttings from trees selected for upright or horizontal branches had branch angles that averaged 18% and 14%, respectively, from horizontal. Repeatability of branch angle was 65%. Cuttings of trees selected for wide or narrow crowns (long or short branches) averaged 29.6 em and 24. 4 em annual branch growth, respectively, and had a repeatability value of 75% (Table 1). Relation to Height Growth The only significant trait to height-growth relation detected was for blue or green foliage. Greenhouse-grown cuttings from green trees averaged 17.4% taller (73. 7 em) than cuttings from blue trees (6 2. 5 em), and field-grown cuttings of trees selected for green foliage (113. 8 em) averaged 16.5% taller than cuttings of trees for blue fohage (97. 7 em) (Table l ). Both differences were signifi­ cant (P < 0. 05) (Table 2). Greenhouse and field data for annual leader growth, cumulative leader growth, tree height, and repeatability for cuttings of each pair of blue or green trees are presented m Table 3. Annual leader growth in 1990 and 1992 and the 1990 to 1992 cumulative growth were all significant in the field test (Table 2). In both environments, only one blue tree produced cuttings that averaged taller than cuttings of the contrasting green tree. Table 1. Number of trees tested, trait repeatability, and tree heights in the greenhouse for four foliage and crown traits. Greenhouse Trait Needles Blue Green Crown Dense Open Branches Upright Horizontal Branches Long Short No. oftrees Repeatability (%) 6 6 66 .3 9 8.3 6 6 89. 2 71.7 6 6 6 5.0 5 5 7 5.0 Field Average phenotype Tree height e( m) 6 2.8 73 .7 I No. oftrees 5 5 Repeatability (%) 90 .1 93.6 Tree height (em) 9 7 .7 I I3 .8 70.9 69 .9 I8 .I o2 14. 2 64 .I 63.7 29.6cm3 24.4cm 73.5 6 5.6 ° 1 Repeatability was the percentage of cuttings of a tree or pair of trees that exhibited the same phenotype as their parent tree(s). Branch angle denoted the average departure in degrees from the horizontal of the longest three branches in the top whorl. Branch length was the average length of the longest three branches in the top whorl. WJAF11(3)1996 79 Table 2 Probability of a > f:.value for leader length and tree height in the field and greenhouse for cuttings of blue and green trees Field Source of variation Blocks ( B ) Pairs Phenotype BxH 1 990 1991 199 2 1990-9 2 Total Tree height Greenhouse tree height 0.0 8 3 0. 73 2 0.034 0. 299 0.05 3 0.3 9 9 0. 13 7 0.5 9 9 0. 208 0.1 29 0.0 26 0.934 0.050 0. 288 0.0 2 2 0.5 4 1 0.04 1 0.0 79 0.040 0.65 3 0.069 0.04 1 0.030 0.606 Table 3. Field1 and greenhouse (G) data for leader growth, tree height, and trait repeatability2 for cuttings of trees selected for blue (b) or green (g) needles. Field Number of cuttings measured Pair Greenhouse Leader growth F G 1990 16 18 1 2 10 14 20 20 20 13 13 4 14 18 19 20 20 20 20 20 20 18 18 20 20 25.0 24.8 20.6 2 7 .I 18.0 2 2 .I 2 1.5 25 .7 16. 2 19.8 Tree 199 2 199 1 Total Tree height/trait repeatability Tree height/trait repeatability Pair mean (em ) (em/%) 2 1.9 71.8 1 16.8/ 100 63.3/ 100 25. 2 84.3/8 9 1 1 6. 2/ 100 7 2.4 19.4 65. 11 100 6 2.6 I01 .9/9 2 26.9 8 2.9 78. 7/100 1 29.4/ 90 16.5 5 1. 2 84.9/100 63.0/100 I 7. 2 1 01.8/100 5 7.9 75.41100 21.4 63 .Oil00 64.4 106.1/ 100 28.I 67. 21100 136.1/100 78. 7 15 .6 70.3/0 46.9 8 2. 1/ 100 18 . 1 5 7.9 93.1/9 2 67.3/100 5 2.1/0 n I sufficient samplel Insufficient comparison sample 68.81 1 00 6 2.8/66 I8.8 97.7/90 58. 7 19.9 23.0 1 13.8/94 73.7/ 98 68.4 2 2.3 (em) lb 2g 3b 2 4g 7b 4 8g 5 9b l Og 13b 7 14g l ib 6 1 2g Average,blue tree Average,green tree 19.9 2 3. 1 24.9 2 2.4 2 2.6 28.9 16.7 18.6 2 1.5 24.9 I5. 1 1 9.9 74 72 69 65 69 60 68 68 Cuttings suppressed by brush or planted in extremely poor microsites were not measured. Repeatability of needle color was the percentage of cuttings that exhibited the same needle color as the parent tree. 3 Field data from trees 11 B and 12G were not analyzed because only four cuttings of tree The genetic control of foliage color was quite strong except in cuttings of two blue trees when grown in the greenhouse. Cuttings of those two trees developed blue foliage in 1990 and 1991, but grew green foliage in 1992. We believe that this irregularity resulted from nitrogen defi­ ciency in 1992 due to our failure to apply fertilize that year. Cuttings of the same clones developed blue foliage in field plots. Selection of phenotypically taller trees in precommercial thinning will continue to be the trait of choice. Precommercial thinning of Sitka spruce in southeast Alaska might enhance height growth of residual and future stands if trees with green foliage also receive preference over trees with blue foliage. Other potential uses of the foliage color and height relation may be in nurseries when seedlings are sorted or in the field when trees are selected for cone collection. One note of caution to blanket endorsement of the above recommenda­ tion is that Faulkner (1987) suggests that in Scotland, seed­ lings from "blue-foliage" populations may be a better choice on frosty or dry sites. 80 WJAF 11(3) 1996 11 B were alive and measurable in 1992. Literature Cited CANNELL, M.G.R.1974. Pinus contorta P and rod u ction of branches and foliage by young trees of Picea sitchensis: simulation. J. Appl. Ec o.l CANNELL, M.G.R. 1982. provenance differences andtheir 11:1091-1115. "Crop" and "isolation" id e oty pe s: Evidence for progeny differences in nursery-grown Picea sitchensis. Silvae Genet 31:60--66. CLARK, J.B., AND G.R. LESTER.1975. The relationship of cuticle structureto the visible ultraviolet sp ectr al pr opertie s of needles from four coniferous species. Plant Physio.l 55:407-413. Rooting Sitka spruce from southeast Alaska. USDA For Serv. Res. N ot e P N W-R N-465. 8 p. FAULKNER, R. 1987. Genetics and breeding of Sitka spruce. Proc. R. Soc Edinburgh93B : 4 1-50. HARRIS, A.S. AND W.A. FARR.197 4. Forest Ecology and Timber Management in The forest ecosystem of southeast Alaska. USDA For. Serv. Gen. Tech Rep. P N W-25. 7 p. SALISBURY, F .B., AND C.W. Ross. 1978. Plantph ys iol og y Wadsworth Pub­ lishing Co., B e lmont, CA. SAS INSTITUTE. 1989. SAS/STAT User's Guide, Version 6. Ed.4. SAS Institute Inc., Ca r y, N C. COPES, D.L.1987. .