New Forests 27: 175–187, 2004.  2004 Kluwer Academic Publishers. Printed in the Netherlands.

Variation in juvenile traits of natural black poplar (Populus nigra L.) clones in Turkey FIKRET ISIK 1, * and FERIT TOPLU 2 1

College of Natural Resources, North Carolina State University, Campus Box 8002, Raleigh, NC 27695, ~ USA; 2 Poplar Research Institute, Ministry of Forestry, PK: 93, 41001 Izmit , Turkey; * Author for correspondence (e-mail: fisik@ unity.ncsu.edu; phone: 1 -919 -515 5029; fax: 1 -919 -515 3169) Received 4 July 2002; accepted in revised form 24 February 2003

Key words: Apical dominance, Clonal variance, Nursery trial, Principal component analysis Abstract. Natural black poplar (Populus nigra L.) clones sampled from river courses were tested in the arid southeast region of Turkey, using a randomized complete block field design with four replications. Clones were laid out in row plots of eight ramets each. Height and apical dominance were assessed at age one year; diameter, survival, bole straightness and branchiness were measured at age two years. Clones differed significantly in survival, growth and quality traits. The results showed that promising clones exist in natural populations. Two commercial clones out of four did not grow as fast as the top new selection clones. For bole straightness, three commercial clones had significantly lower grading scores than the top best 14 clones. Principal components analysis indicated that growth, apical dominance and branching are the most important traits distinguishing black poplar clones. Diameter had a moderate correlation (0.34) with bole straightness. Relationships between geographic variables (elevation, latitude and longitude) with growth and stem quality traits were weak. Considerable genetic variation was observed among clones for all the traits. Genetic differences among the clones accounted for 27% (survival, bole straightness) to 39% (height) of the total variance. Broad-sense individual heritability ranged from 0.27 (survival) to 0.37 (apical dominance). Clonal mean heritabilities were higher than individual heritabilities and ranged from 0.60 (survival) to 0.82 (diameter), implying considerable gain could be realized via selective improvement methods.

Introduction Black poplar (Populus nigra L.) is native to the Asian part of Turkey, particularly in the inland and eastern parts (Tunc¸taner 1995). The species is most often found along rivers. Natural populations of the species are under risk of extinction due to their scattered natural occurrence and increasing human pressure. In Turkey, black poplar shows pyramidal crowns with two distinct forms. One type is Populus nigra cv. Italica mainly found in western Anatolia. The older individuals of this form have black or dark greyish bark. While the second form, Populus usbekistanica Komarov cv. ‘Afghanica’, has thin, smooth, and white bark (FAO, Rome, Italy 1979; Tunc¸taner 1995). Cultivars and hybrids of the species are widely used in plantations in Turkey. The species has a significant economic value due to its wide geographic adaptation and fast growth. According to a poplar wood survey, black poplar comprised 57% of the

ICPC - XPS 110866 (NEFO) - product element NEFO209-02 - Fri Jun 13 14:11:46 2003

176 annual poplar wood production in Turkey, which was about 3.5 million m 3 (Birler and Koc¸er 1992). Most of the black poplar wood harvested in Turkey is used for construction and making boxes. Peeled unprocessed poles are commonly used for rafters, posts and beams. Poplar is the main source of construction wood for local markets, particularly in semi-arid and barren regions of the country. If certain bole traits (diameter growth, straightness, less brachiness) and crown traits (narrow crown) were genetically improved, then there would be other potential uses of processed wood including veneer, plywood and furniture. Black poplar is often used as a male parent in hybrid crosses. Mating between American Populus deltoides Bartr. and Asian-European black poplars is highly successful, particularly when black poplar is used as a male parent. The progenies of these hybrids can be easily propagated by cuttings. Investigations on hybridization between P. deltoids and P. nigra indicated that improvement on Melampsora rust disease and Marssonina brunnea can be accomplished (FAO, Rome, Italy 1979). Wood demand in Turkey is expected to increase, particularly in the semi-arid Southeast Anatolia Region (GAP) region where a large irrigation project has been carried out. At completion, 1.7 million hectares of land will be irrigated. As a result, agricultural production (cotton, various vegetables and fruits) is estimated to double to 16 million tons per year. Packaging, storing, and transporting agricultural products will ultimately increase wood demand in the region by 1.5 million m 3 annually (Birler and Koc¸er 1992). Poplar plantations seem to be the most efficient way to meet this wood demand. According to a survey, if 5% of the irrigated land in the region is reserved for poplar plantations, about 2.5 million m 3 wood could be produced annually (Birler and Koc¸er 1992). Poplar plantations have acquired a considerable economic importance in many regions of the world, including southern Europe, Middle East and North America (FAO, Rome, Italy 1979). Poplars exhibit more rapid early growth than conifers (Laing et al. 1985). Wide natural distribution, considerable genetic variation and relative ease of hybridization and vegetative propagation have made poplars attractive for genetic improvement and plantation programs (Brown et al. 1996). Short rotation ages and easy planting are other advantages of poplar cultivation. For solid wood production, the rotation age of black poplar plantations ranges between 10 to 20 yr in Turkey. Rotation for pulp production can be as short as four years (Heilman 1999). A program was initiated in Turkey to conserve and improve black poplar genetic resources in 1994 (Tunc¸taner 1993, 1995; Toplu 1996). Within the framework of the program, clonal trials were established in various regions. Information on genetic variation is crucial in conservation and improvement of tree species. The magnitude of genetic gain through clonal selection depends on variation and inheritance of desired traits (Foster 1986). There is still lack of information on the structure of genetic variation of natural black poplar populations that occur in Turkey. In this study, variation in some juvenile traits of black poplar clones from a two-year-old nursery trial located in the GAP region are reported. The objectives of the study were to: (i) estimate genetic variances among and within clones (ii) investigate genetic and morphological traits that distinguish natural clones and (iii) assess black poplar clonal performance in a semi-arid region of Turkey.

177 Materials and methods

Clonal selection A survey was carried out including all major river valleys in Turkey to locate natural stands of black poplar (Figure 1). In addition, plantations over 50 yr of age were identified in the surveyed regions. Several commercial black poplar clones have been available in the market for about 45 yr. In order to avoid duplication in sampling commercial clones, the history of plantations was also taken into account, i.e., by interviewing the owner about the source of cuttings. Each river valley or its major tributary was taken as a selection unit (population). In case of a long river, more than one selection unit was sampled. The distance between two selection units in the same river valley was about 50 km. One female and one male parent were selected from each population for future hybridization and selection programs. Primary selection criteria were age (. 50 yr of age), growth, bole straightness and apparent health. In addition to selecting one male and female individual from a selection unit, trees having obvious differences from the selected ones in bark morphology, bole straightness, crown form and leaf characteristics were also sampled. Altogether, about 300 individuals were selected throughout Turkey. From each clone, a 70 to 100 cm long branch was collected from the upper crown, labeled and kept in a damp cloth until transferred to the Poplar Research Institute at Izmit (Figure 1). Each branch was cut into a few cuttings that were then rooted in a nursery to produce more cuttings for field planting. Due to limited number of cuttings, only 236 clones, including four commercial check lots were transferred to a nursery trial. The commercial clones (Clone [ 1, 2, 12 and, 13) were used as check lots (controls) to compare with the new selected material.

Figure 1. Locations of black poplar material collection and test site in Turkey.

178 Site and experimental design The experiment was established in S¸anlıurfa Harran Rural Affairs Research Station (altitude 530 m) in the southeastern part of Turkey. The region is one of the driest parts of the country with 388 mm annual rainfall. The mean annual temperature is 17.1 8C. Summers are dry and can be very hot. The absolute maximum temperature in summer can reach 46.6 8C (July) and the absolute minimum temperatures can be 216.8 8C (January) in winter (Anonymous 1998). The soil texture is a clay loam; pH is about 7.8, with salt content 0.07 milimhos / cm but is fertile with total 25.2% of CaCo 3 . The site was ploughed and disked two times in 1996. The size of the cuttings was about 20–22 cm in length and about uniform in thickness. Standard planting techniques were followed, i.e., burying the scion completely in the soil and only leaving the tip at the surface. The experimental design was a randomized complete block with four replications of 8-tree clonal row plots. Thus, initially 1024 unrooted cuttings (236 clones 3 8 ramets 3 4 replications) were planted at the experiment site. The spacing between rows was 220 cm and between cuttings in a row was 40 cm. The experiment was established in February 1997. Starting from mid-May, the site was irrigated at two-week intervals until end of September both in 1997 and 1998. Weed control was carried out three times a year by a tractor-drawn rolling cultivar. Assessment of the traits When the saplings were one-year-old, height was measured to the nearest cm in January 1998. To assess branchiness of the clones, a branch index was derived. The difference between sapling height and the height of the third main branch from the top of a sapling was first calculated. Then, this value was divided by sapling height. Clones with branch index close to 1.0 have fewer branches and are desirable. In the first year, survival of the clones was also determined. Observations in the first year were done on a subsample of the four middle saplings out of eight due to limitations in labor and time. In the second year, diameter at 1.0 m above ground level was measured to the nearest mm. Apical dominance of terminal shoots was assessed visually using a five point score, where 5 is the best (one dominant shoot) and 1 the worst (several dominant shoots). Similarly, bole straightness was assessed visually using a four point grading score, where 4 was the straightest. Second year observations were carried out on six saplings for each clone in a given plot, allowing one border tree at the beginning and at the end of the row. The timing of bud burst and leaf shedding were assessed visually using a four point and five point score, respectively. Full bud burst and complete leaf shedding were assigned 4 and 5 scores, respectively. The observations were carried out on a weekly basis starting from March 8 th for bud burst and October 30 th for leaf shedding in 1998. Statistical analyses Analyses of variance were conducted on individual values for all the traits to detect

179 significant differences among the clones, except survival, applying the following linear model: y ijk 5m 1b i 1c j 1bc ij 1e ijk Where, y ijk observed kth sapling phenotypic value in the ith block of jth clone, m is the overall mean, b i the random effect of block i, i 5 1 . . . 4, c j the random effect of clone j, j 5 1 . . . 236, bc ij the random interaction effect of jth clone in the ith block, and e ijk the random sapling error of kth sapling in jth clone in the ith block, assuming NID (0, s e2 ). Survival was determined as the number of living saplings in a plot over the initial number of planted cuttings. Arcsine-transformed values were used in the analysis to fulfill analysis of variance assumptions (Sokal and Rohlf 1995). For survival, a reduced linear model was applied in the analysis of variance, omitting the blockclone interaction term. Least squares means of clones were estimated using the LSMEANS option in the GLM procedure of SAS (SAS Institute Inc. 1989). Variance components were estimated by applying the SAS VARCOMP procedure and using the Restricted Maximum Likelihood (REML) method (Table 1). Broadsense individual (H 2i ) and clonal mean heritabilities (H c2 ) were estimated using the following equations:

s 2c 2 H i 5]]]] s 2c 1s 2bc 1s 2e s c2 2 H c 5]]]] 2 s bc s e2 s c2 1]1] b bn Where s 2c is genetic variance due to clones, s 2bc is block-clone interaction (plot) variance, s 2e is within plot variance, b is number of blocks and n is number of saplings per clone per block. Principal component analysis was applied to interpret traits that separate clones using the clonal means and applying PRINCOMP procedure of SAS. Product-moment correlations between growth and geographic variables (i.e. latitude, longitude and elevation) of the clones were detected using CORR procedure of SAS Institute Inc. (1989).

Results Variation among clones There were highly significant (p , 0.0001) differences among clones for diameter,

180 Table 1. Two-way analysis of variance model and expected mean squares equations used for the black poplar clonal trial established in Harran region, southeast Turkey Source Blocks Clones Block-clone Error

D.F b-153 c-15235 (b-1)(c-1)5702 bc(n-1)53776

Expected Mean Squares 2 e 2 e 2 e 2 e

2 bc 2 bc 2 bc

s 1k 4 s 1k 5 s s 1k 2 s 1k 3 s s 1k 1 s s

2 b 2 c

F statistics EMS b / EMS bc EMS c / EMS bc EMS bc / EMS e

Number of individuals observed within each clone n ranged from 16 to 24 depending on the character. D.F. 5 Degrees of freedom, s 2e 5 Residual variance, s 2cb 5 Block-clone interaction variance, s 2c 5 Variance due to clones, s 2b 5 Variance due to blocks. k 1 , . . . k 5 , variance components coefficients (k 3 515.8 for height and branch index, k 3 523.16 for diameter, apical dominance and bole straightness).

height, bole straightness, survival and branchiness (Table 2). The results showed that superior clones exist in the natural populations. For diameter growth, the top nine clones were not significantly different from each other. Commercial hybrid clone 12 had the greatest height and diameter growth, followed by new selection clone 28 (Figure 2). Commercial clone 2 was also among the top 9 clones for diameter growth but two other commercial clones, 13 and 1, were not. For bole straightness, three commercial clones (1, 2 and 13) out of four had significantly lower straightness scores than the top clones. Commercial clones 1, 2 and 13 ranked 76 th , 26 th and 28 th , for bole straightness, respectively. A considerable proportion of the phenotypic variance for the traits studied was

Figure 2. The top 14 clonal least squares means (6 1 standard error of mean) for diameter growth at age two years. The shaded clones are commercial hybrids, whereas the solid white bars are new selected clones. The horizontal line on the bars shows that the first 9 clones from left are not significantly different.

181 Table 2. Mean squares (MS) and F statistic probability levels (Pr .F) for the factors in the linear model for height, diameter, branching index (BRI), bole straightness, apical dominance and survival Source

Height MS Pr.F

Diameter MS Pr.F

MS

BRI Pr.F

Straightness MS Pr.F

Apical D. MS Pr.F

Block 193358 ,.0001 742.74 ,.0001 0.0496 0.3819 4.751 0.0018 Clone 18769 ,.0001 337.15 ,.0001 0.1844 ,.0001 3.149 ,.0001 Plot 3890 ,.0001 62.13 ,.0001 0.0489 ,.0001 0.955 ,.0001 Error 432 15.20 0.0090 0.113

Survival MS Pr.F

35.541 ,.0001 0.0170 0.352 9.252 ,.0001 0.0390 ,.0001 2.224 ,.0001 0.157 0.0156

Probability levels Pr,0.05 are significant. Table 3. Variance component, variance (%) component proportion over total variance, individual (H 2i ) and clonal mean (H 2c ) broad-sense heritabilities for the traits studied Height

Diameter

BRI

Straightness

Apical D.

Survival

Source Estimate

% Estimate

% Estimate

% Estimate

% Estimate

% Estimate

%

Block Clone Plot Error

202.8 948.8 870.3 432.4

8 0.524 39 12.024 35 8.014 18 15.193

2 34 22 42

0 32 36 32

1 27 40 32

3 36 42 19

0 27 – 73

H 2i H 2c

0.3460.027 0.8060.020

0.3460.027 0.8260.019

0.0 0.0091 0.0102 0.0090

0.3260.030 0.7560.027

0.0027 0.0965 0.1436 0.1132

0.2760.029 0.7060.032

0.0249 0.3021 0.3519 0.1569

0.3760.031 0.7660.025

0.0 0.00585 – 0.01558

0.2760.035 0.6060.043

genetic. Genetic differences among the clones explained 27% (survival) to 39% (height) of the total variance (Table 3). High genetic variances due to clones were reflected in high broad-sense heritabilities (Table 3). Broad-sense individual heritabilities of the traits ranged from 0.27 (survival) to 0.37 (apical dominance). Broad-sense clonal mean heritabilities were higher than individual heritabilities and ranged from 0.60 (survival) to 0.82 (diameter). Height was positively correlated with branch index (r 5 0.20) (Table 4). In contrast, the relationship between branching and diameter was weak (r 5 0.04). Diameter was positively correlated with bole straightness (r 5 0.34), which is desirable. Height and diameter were significantly correlated with apical dominance (r 5 0.60 and r 5 0.72, respectively). None of the observed characteristics were correlated with the original altitudes of the clones (Table 5). Although height, diameter and apical dominance had significant correlations with longitude, the relationships were weak and ranged from r 5 20.03 to r 5 20.16. Principle component analysis Eigenvalues indicated that the first three components explain 90% of the overall variation in the data (Table 6). Scatter plots of principal components illustrated differences among clones. The distribution of clonal means in two-dimensional space was not discrete but overlapping (Figure 3). The first component (PRIN1) explained 52% of the variance. In the first component, height, diameter and apical dominance had approximately equal and positive loadings. Bole straightness had a positive but comparatively smaller loading. Clones with large scores of PRIN1 were

182 Table 4. Product-moment correlation coefficients among traits for black poplar clonal means

Diameter Branch index Apical dominance Bole straightness Survival Bud set Leaf shedding

Height

Diameter

0.84 0.20 0.60 0.17 0.35 0.10 0.44

0.04 0.72 0.34 0.34 0.03 0.49

Branch index

Apical Dom.

Bole Straight.

Bud Set

0.12 20.24 20.26 0.21 0.10

0.35 0.32 0.05 0.48

0.34 20.09 0.30

0.02

Correlation coefficients are significant r . 0.13 at 0.05 level and r $ 0.17 at 0.01 level. Number of clonal means used in the analysis ranged from N 5 224 to N 5 236. Table 5. Correlation coefficients among geographic variables and measured traits of the black poplar clonal means Height Latitude 20.05 Longitude 20.13 Altitude 0.03

Diameter 20.07 20.16 20.01

Branch Index 0.16 20.07 20.12

Apical dominance 0.07 20.18 20.09

Bole Straight. 20.09 20.07 20.17

Bud Burst 0.09 20.03 20.01

Leaf Shedding 0.11 20.13 20.06

Correlation coefficients are significant r . 0.13 at 0.05 level and r $ 0.17 at 0.01 level. Number of clonal means used in the analysis ranged from N 5 224 to N 5 236.

good performers and had higher apical dominance (Figure 3), whereas clones with smaller scores were inferior to the site mean for growth traits. For example, one of the fastest growing clone [12 had a PRIN1 value of 2.17, whereas a slow growing clone [60 had a negative value (22.33). The second component (PRIN2) was primarily related to branchiness and bole straightness, and explained 25% of the standardized variance. Clones with higher scores of PRIN2 had higher branching index (less branchiness) but poorer bole straightness compared to clones having low PRIN2 values. For example, clone [67 had one of the highest branching indexes (0.70) and had the highest PRIN2 scores (Figure 3 and 4). The same clone had also one of the lowest bole straightness score (1.5 out of 5). The third component (PRIN3), explaining 13% of the variance was again related to bole straightness and branching index but with positive loadings for both traits. Large scores of PRIN3 identified clones with straighter stems and less branching (Figure 5). Clone [119 for example, had the most desirable stem quality as shown by higher PRIN2 and PRIN3 values (Figure 5). This clone had one of the greatest bole straightness scores (3.0) and branching indices (0.7). Factor scores of PRIN4 were mainly related to apical dominance and had a negative loading. However, this component explained only 8% of the variance. Collectively, subsequent components contributed less than 2% of the variance.

Discussion A basic understanding of genetic structure of natural populations of a species is

183

Figure 3. Scatter plot of principal components Prin1 (growth) and Prin2 (branchiness) of black poplar clones. Clones having high positive Prin1 and high positive Prin2 values (e.g. 167, 132) had greater height and fewer branchiness.

Figure 4. Scatter plot of Prin1 (growth) and Prin3 (bole straightness, branching index) of principal components of black poplar clones. Larger Prin1 values indicate faster growth, whereas, larger Prin3 values (e.g. clone 119, 141 67) indicate straighter bole and fewer branchiness.

184

Figure 5. Scatter plots of PRIN2 (branchiness) and PRIN3 (bole straightness) principal components of black poplar clones. Clones 119, 141 and 152 had the most desirable quality, whereas, clones in the lower left region had undesirable branching characteristics. Table 6. Results of principal component analysis of black poplar clones Traits Height Diameter Branch index Apical dominance Bole straightness Eigenvalue Variation explained (%) Cumulative (%)

PRIN1 0.5398 0.5846 0.0729 0.5310 0.2822 2.6012 52 52

PRIN2

PRIN3

PRIN4

0.2164 2.0066 0.7616 0.0073 2.6108 1.2641 25 77

2.3065 2.2281 0.6229 0.1200 0.6720 0.6343 13 90

0.4609 0.1539 0.1176 2.8129 0.2986 0.3818 8 98

crucial before implementing a breeding and gene conservation program (Foster 1986). Casual variance components and broad-sense heritabilities estimated in our study suggested that there is considerable genetic variation among natural black poplar clones in Turkey. Selective methods would result in considerable genetic gain from clonal selection for growth and quality traits. Selection based on clonal means would be more efficient than selection on individual trees because clonal mean heritabilities were much higher than individual tree heritabilities. However, the heritabilities may be biased upward due to possible genotype-environment interaction effects that were not separable. In addition, clonal genetic variance and heritabilities in this study could be inflated because of ‘C’ effects. C effects include common environmental factors associated with the cutting size, the location of the cutting on the tree and rooting environment differences (Libby and Jund 1962;

185 Stonecypher and McCullough 1986). Significant C effects were reported for rooting ability for Tsuga heterophylla (Raf.) Sarg. (Foster et al. 1985). None of the sapling characteristics was significantly correlated with the altitude of the collection sites. This is partly due to the nature of sampling employed, which was based on river courses as a selection unit rather than following geographical variables. Similarly, weak relationships among latitude and longitude origin and growth and quality traits may be explained by the sampling pattern. Principal components analysis results indicated that growth and branch index were the most important characters that separated clones among the traits studied. Similar to observed traits, the correlations among factor scores of the principal components and geographic variables were weak and could also be due to a restricted distribution of the species that follows rivers but not geographic gradients. These results are in accordance with a similar study carried out by Cagelli and Bisoffi (1994) on Populus nigra clones. They concluded that two main components explained more than 50% of the total variation among the black poplar clones. The first component was linked with growth, survival and bole straightness, whereas, the second component was related to the branch index. Our study showed that promising clones from natural poplar plantations exist. Seven clones selected form natural stands grew as fast as commercially available clones [12 and [2 at age two. These seven new clones outperformed control clones [1 and [13 for diameter at age two (Figure 2). The superiority of these seven new selected clones over the commercial clones for bole straightness was even greater (Figure 3). Utilization of genetic variation from the natural provenances of poplars has been largely neglected (Foster 1986). However, if proper natural sources are selected within a species, the cheapest and fastest gain can be realized (Zobel and Talbert 1984). These results suggested that a combination of selection from natural stands with hybridization strategy might serve to best secure long-term genetic gain and conservation of black poplar natural clones. Clonal selection is a stepwise process of evaluation. It starts with a large number of genetically different progenies and ends up with a few clones with commercial utility. Generally, about the best 10% of total clones are selected for the first phase (Cagelli and Bisoffi 1994). Ares (2002) reported high age-age correlations for growth traits and disease incidence from black poplar clonal trials in Argentina. However, performance of the clones may change significantly by age, particularly for disease incidence, bole straightness and taper (Steenackers et al. 1993). The temptation to make extreme selections in the seedling nursery should be limited to culling crooked plants or those that are found to be susceptible to diseases. Growth during the first season can be strongly affected by the time of transplanting, and can, thus not be an efficient selection criterion. Brown et al. (1996) reported that first year height growth in the field and in a control environment was not a good predictor of ranking of clones after five growing seasons of several poplar hybrids. Although relatively uniform material was used in this study, the final selection of the clones should be long enough to assess performance of clones in this semi-arid region of Turkey. At this stage, a small percentage (e.g., 10%) of poor performers and clones prone to any disease can be excluded from long-term trials. These trials

186 should be replicated across potential poplar cultivation areas throughout the region. A half of the rotation age has been suggested for the final selection of introduced species to reduce risks of maladaptation (Zobel and Talbert 1984; Lo et al. 1995). Long-term trials also provide information to validate growth and disease infection response of clones. The region where the test was established experiences severe droughts compared to original location of most natural populations of black poplar. Clones from Yozgat and Tokat regions (e.g., clone [60, [67) may have adapted to a more humid environment compared to GAP region. These populations were poor performers at the test site. Most populations from the north and west of Turkey can be considered ‘exotic’ for the region. Thus, for the final selection, clones should be tested for a minimum of 15 yr of age on replicated sites before final selection for commercial utility.

Conclusions There were significant differences among the clones for growth, survival and quality traits. Growth and branch index were the most significant characters that separated the clones. Several clones from natural populations had greater growth and better quality than commercial clones. High broad-sense heritability for height, diameter and quality traits suggested that a considerable proportion of the variation is genetic. Natural black poplar populations have good potential for improvement of growth and quality traits. Significant improvement could be realized by applying selective tree improvement techniques on black poplar. A strategy combination of selection from natural populations with hybridization may give higher genetic gain than hybridization alone. Including clones from natural populations may better serve long-term genetic gain and conservation of genetic resources of black poplar. The natural stands of black poplar represent a diverse genetic variation of the species in Turkey. It is therefore important to conserve natural populations. A clonal archive should be established for ex situ conservation of natural black poplar genetic resources. A minimum of 15 yr of testing should be considered for final selection of clones for commercial utility.

Acknowledgements We are grateful to The Scientific and Technical Research Council of Turkey ¨ ~ (TUBITAK) for financial supporting of the project. We are also thankful to Mr. ¨ ¸tu¨ Ertas for their close interest and invaluable help in Sabahattin C¸elik and Rus establishment and maintenance of the experiment at Harran Research Station of Rural Affairs Research Institute. We also thank three anonymous referees for their valuable suggestions and contributions in improving the manuscript.

187 References Anonymous 1998. Turkish Rural Affairs Hydro Meteorological Observations Annual Report, Harran, S¸anlıurfa, Turkey. Ares A. 2002. Changes through time in traits of poplar clones in selection trials. New Forests 23: 105–119. Birler A.S. and Koc¸er S., Ministry of Forestry of Turkey, Poplar Research Institute 1992. The socioeconomic importance of poplar cultivation in Southwest Anatolia Region of Turkey. Miscellaneous publications No. 1. Izmit, Turkey, 26 p. Brown K.R., Beal F.D. and Hogan G.D. 1996. Establishment-year height growth in hybrid poplars; relations with longer-term growth. New Forests 12: 175–186. Cagelli L. and Bisoffi S. 1994. Variability of juvenile traits within the Italian collection ofPopulus nigra and selection of superior clones. In: Frison E., Lefevre F., de Vries S. and Turok J. (eds), Populus ~ nigra Network. Report of the first meeting, 3–5 October 1994, Izmit, Turkey. IPGRI, Rome, Italy 20–21. FAO, Rome, Italy 1979. Poplar and Willows in Wood Production and Land Use. FAO Forestry Series No. 10. p. 34–38. Foster G.S. 1986. Provenance variation of western cottonwood in the lower Mississippi valley. Silvae Genetica 35: 32–38. Foster G.S., Campbell R.K. and Adams W.T. 1985. Clonal selection prospects in western hemlock combining rooting traits with juvenile height growth. Can. J. For. Res. 15: 488–493. Heilman P.E. 1999. Planted forests: poplars. New Forests 17: 89–93. Laing F.M., Sendak P.E. and Aleong J. 1985. Species trials for biomass production on abandoned farmland. North. J. Appl. Forestry. 2: 43–47. Libby W.J. and Jund E. 1962. Variance associated with cloning. Heredity 17: 533–540. Lo M.H., Abrahamson L.P., White E.H. and Manion P.D. 1995. Early measures of basal area and canker disease predict growth potential of some hybrid poplar clones. Can. J. For. Res. 25: 1113–1118. SAS Institute Inc. 1989. SAS / STAT User’s Guide. 4th edn. SAS Institute Inc., Cary, NC, 846 p. 2, Ref: version 6. Sokal R.R. and Rohlf F.J. 1995. Biometry. 3rd edn. W.H Freeman and Company, New York, 887 p. Steenackers J., Steenackers V., Van Acker J. and Stevens M. 1993. Stem form, volume and dry matter production in a twelve-year-old circular Nelder plantation of Populus trichocharpa x deltoides Beaupre. For. Cron. 69: 730–735. Stonecypher R.W. and McCullough R.B. 1986. Estimates of additive and non-additive genetic variances from a clonal diallel of Douglas-fir Pseudotsuga mensiesii (Mirb.) Franco. P. 211–227 in IUFRO Conference. A joint meeting of working parties on breeding theory, progeny testing, seed orchards. Conference proceedings. October 13–17, Williamsburg, VA. Toplu F. 1996. Breeding and conservation of genetic resources of black poplar (Populus nigra L.) in Turkey. 20 th Session of the International Poplar Commission. Budapest, Hungary, 1–4 October 1996. p 623. Tunc¸taner K. 1993. Conservation and utilization of poplar and willow genetic resources in Turkey. ~ Publication No. 3. Poplar Research Institute, Izmit, Turkey, 23 p. Tunc¸taner K. 1995. Conservation of genetic resources of Populus nigrain Turkey. In: Frison E., Lefevre F., de Vries S. and Turok J. (eds), Populus nigra network. Report of the first meeting, 3–5 October ~ 1994, Izmit, Turkey. IPGRI, Rome, Italy. Zobel B. and Talbert J. 1984. Applied Forest Tree Improvement. John Wiley Sons, 505 p.