Growth and Phenology of Hybrid Aspen Clones ( Populus tremula L . × Populus tremuloides Michx . )

Height, basal diameter, diameter at breast height, bud burst, and leaf development were recorded in a 5-year-old hybrid aspen clonal trial. The fi eld trial consisted of four aspen hybrid clones (Populus tremula × Populus tremuloides ) and one local P. tremula seedling source. Phenological traits were observed in the 3rd year. Growth patterns were recorded during the 3rd and 4th years. Phenological traits were explored in relation to hybrid vigor expressed as growth traits. Differences were observed for phenological and growth traits among hybrid clones and P. tremula. The growth period varied from 143–158 days for the four hybrid clones, and was 112 days for P. tremula. The correlation between growth period and yield was highly signifi cant. The annual growth rate of height for the hybrids was 4.2 cm per 7 days (2.4 for P. tremula) in the 3rd year and 6.4 cm per 7 days (2.9 for P. tremula) in the 4th year. After 5 years, mean estimated stem volume of the hybrids was 3.9 times that of P. tremula. Signifi cant clone by year interaction was observed for height, diameter, and volume growth. The hybrid vigor seems to be mainly attributable to a longer growth period.


Introduction
Populus tremula is a widely distributed species in Finland, often growing in mixed stands with pine, spruce and birch.Recent developments in wood fi bre research have allowed short fi bres to be used in admixture with coniferous long fi bres in high-quality papermaking.The Finnish forest industry has shown a fresh interest in utilising aspen for producing the short fi bres.Interest in cultivating selected clones of hybrid aspen has increased.The hybrid between European aspen P. tremula L. and North American trembling aspen P. tremuloides Michx.has shown superior growth Silva Fennica 35(1) research articles in Finland (Beuker 1989).Genetic gain from aspen breeding programs were demonstrated in the United States (Einspahr 1984) and Europe (Melchior 1985).Interspecifi c hybrids have grown faster than the progenies of intraspecifi c crosses (Heimburger 1936, Einspahr and Benson 1964, Melchior and Seitz 1966, Heimburger 1968, Zsuffa 1969, 1973).There is much genetic diversity in aspen, and it can be manipulated through selective breeding, interspecifi c hybridization and cloning (Li 1995).
Hybrid vigor has been observed in poplar and has been one of the driving forces in poplar breeding (Muhle Larsen 1970).The phenomenon is also well documented for aspen hybridisation in the U.S. (Li and Wu 1996, 1997, Li et al. 1998).This increase in vigor may have many components, including carbon allocation patterns, water and nutrient use effi ciency, and shoot growth phenology.Traits that affect Populus performance include phenology (Michael et al. 1990, Ceulemans 1992), leaf and stomatal morphology and photosynthetic capacity.Large clonal differences in phenology, leaf area development, and photosynthetic production have been found (Ceulemans 1987, Orlovic 1998, Thomas et al. 1997).Li et al. (1998) found that interspecifi c aspen hybrids grew faster than intraspecifi c hybrids at the juvenile stage.The authors ascribe this to larger internode number and length and leaf number.Heterosis of P. tremula × P. tremuloides hybrids for stem volume was probably the result of delayed bud set, which gave a longer duration of height growth.
In this paper we compare hybrid aspen clones and local P. tremula for yield, growth pattern and phenology.The aim of the study was to obtain better fundamental understanding of hybrid vigor of aspen hybrids.We present results of studies on yield components and their relationship in hybrid aspen clones and local P. tremula.The growth patterns were studied mainly from a phenological perspective, in an attempt what has hitherto been loosely termed heterosis of species hybrids.

Materials and Methods
The material used in this study was a clonal trial established by the Foundation for Forest Tree Breeding and the University of Helsinki.The trial is located in Viikki, area of Helsinki University (lat.60°14´ N, long.25°05´ E, alt 10 m).The trial was planted in May, 1994 with one-year-old plants, and consists of four aspen hybrid clones (Populus tremula × P. tremuloides) and one P. tremula seedling source.Clone 1 was a hybrid between P. tremula (Tuusula, south Finland) × P. tremuloides (Lake Aleza, B.C. Canada).Clone 2 was a hybrid between P. tremula (Tuusula, Finland) × P. tremuloides (Maple, Ontario, Canada).Clones 3 and 4 were selected from a mixed plantation of hybrids, between P. tremula (Tuusula, or Punkaharju, Finland) and P. tremuloides (Lake Aleza or Maple, Canada).Thus their exact pedigrees are unknown.The local aspen seedlings were progenies of single open pollinated superior tree in Karkkila (lat.60°32´ N, long.24°15´ E), Finland.Thus they are predominantly half-sib.A randomized complete block design was used with 4 × 4 plants spaced at 2.5 m × 2.25 m per plot and fi ve blocks.
Total height, basal diameter 15 cm from the ground and breast height diameter at height 1.3 m were measured.Measuring positions were marked on the stems to assure that repeated measurements were taken at exactly the same position.Stem volume was estimated as height × basal diameter 2 (Causton 1985, Ceulemans 1992, Li et al. 1998).
Height, basal diameter and breast height diameter were measured in the autumns of the 2nd to the 5th year.For observation of dynamic seasonal growth patterns, the fi ve tallest trees were sampled for measurements of growth traits in each plot in the spring of the 3rd year.Height, basal diameter and breast height diameter were measured at intervals of approx. 2 to 3 weeks over the course of the growing season in the 3rd year (10 measurements) and the 4th year (8 measurements).The growth rates for height, basal diameter and breast diameter were expressed as increments (cm or mm) over a 7 day period.The annual growth rates were calculated by bringing total growth made over the total growth measuring period (days) to a 7 day basis.
Phenological traits were recorded in the 3rd year.Growth initiation was classifi ed into three stages counting from the fi rst of May: 1) the day when the bud emerged 5 mm beyond the bud scale, 2) the day leaves unfolded, 3) the day the leaves attained full size.Likewise, three stages of growth cessation were distinguished: 1) the day when half of the leaves yellowed, 2) when all of the leaves yellowed, 3) when all leaves defoliated."Bud emerged" and "half of the leaves yellowed" were taken as indicators of growth initiation and growth cessation.The length of the growth period was estimated as the number of days from growth initiation to growth cessation.Each individual was observed from the same direction and from about the same relative position with respect to the tree crown, every other day in the spring, and every third day in the autumn.
All calculations were based on plot means.The difference between the hybrid clones and the P. tremula were established by an analysis of variance (ANOVA) using the PROC GLM procedure of the SAS statistical software package (1989; SAS Institute, Cary, NC) with the type III estimation of sum of squares.Means of phenological and growth traits were separated by Tukey's multiple range test at p ≤ 0.05 probability level of signifi cance.Pearson correlations were calculated to assess the linear relationships between the studied traits.Data from all the four years were combined and the following statistical model was used: where X ijkl is an observation on the l th tree from the i th clone in j th year and k th block, µ is the overall mean, C i is the effect due to the i th clone, Y j is the effect due to the j th year, B k(j) is the effect due to the k th block in the j th year, C i Y j is the interaction of the i th clone and j th year, is the interaction of the i th clone and k th block in the j th year and ε ijkl is random error.Clonal effects were considered to be fi xed, while the year and block were treated as random effects.

Growth
Highly signifi cant differences in growth traits (height, basal and stem volume) were observed among the blocks, clones, years and their interactions (Table 1).All hybrid clones had outperformed the P. tremula by the end of the 5th year.The best clones were clone 2 and 1, with stem production volumes of 17.82 dm 3 and 17.57 dm 3 , respectively (Fig. 1).Clonal ranking in the 5th year remained nearly the same as in the 2nd year.Clone 3 was shorter than clone 4 at the end of the 5th year, but because of its larger diameter, it had a larger stem volume (Table 2).
Growth patterns of mean height, basal and breast diameter of the hybrid clones and P. tremula in the 3rd and the 4th year are shown in Fig. 2. The hybrids displayed superior growth for all traits during the entire growth period (Table 3).
The growth rate of height, basal and breast height diameter was highest from the beginning of June to the end of July in the 3rd year.The same pattern for the basal and breast height diameter was found in the 4th year.However, for height growth, the highest growth rate occurred during mid August.The growth rate of hybrid aspen clones had two peaks, the fi rst in mid-July, and the second in mid-August.For the local P. a Sums of squares are multiplied by 10 -3 tremula, however, there was only one peak of growth in mid-June.(Fig. 2).The growth patterns were different for different clones.Even for the same clone, the growth patterns for basal and breast diameter differed from the pattern for height growth.Clone 3 had the highest growth rate for basal diameter in mid-July in the 3rd year, but ranked third for height and breast height diameter (Fig. 2).
The growth patterns in the 3rd and the 4th year were different (Fig. 2).Compared with hybrid clones in the 3rd year, the annual growth rate of the P. tremula was 56% for height, 73% for basal diameter and 71% for breast height diameter.The respective values in the 4th year were 45%, 67% and 56%, respectively (Table 3a, 3b).

Phenology
Hybrids fl ushed earlier and grew longer than P. tremula (Table 4).The mean time for each stage of growth initiation of the hybrids was earlier than for P. tremula: 9 days for bud emerged, 7 days for leaves unfold, 7 days for full size leaves.The ranges of variation of growth cessation varied from clone 3 to the P. tremula : 4-41 days for half leaves yellowed, 11-45 days for all leaves yellowed, and 16-54 days for all leaves defoliated.For the hybrids, averaged half leaves yellowed, all leaves yellowed and all leaves defoliated were all 32 days later than for P. tremula.Hybrids had a growth period of between 158-143 days compared to 112 for P. tremula.The mean of the growth period of hybrids was 41 days   longer than for P. tremula, the four hybrids all fell within a 15 day range.

Correlations
Signifi cant correlations were observed among phenological and growth characters.The three stages of growth initiation were negatively correlated with stages of growth cessation.Growth traits were positively correlated with growth cessation and growth period, but negatively correlated with growth initiation (Table 5).This means that late starters generally stopped grow-  Note: Means in vertical sequence not followed by the same letter are signifi cantly different at P < 0.05 according to Tukey's test.
ing early, and thus had a short growth period.This must be ascribed to the different parental origins of the F1 hybrids.

Discussion
In our study, the superiority of the hybrid clones was best illustrated by stem volume.Hybrid clones had a signifi cantly greater wood production than P. tremula.In spite of obvious superiority in hybrid yield, hybrids varied considerable for stem volume and annual growth rate.The hybrid clones showed signifi cantly higher growth rates and yields than P. tremula at most observation dates over the entire growth season in the 3rd and 4th years (Table 3).Generally, the higher the yield, the higher the growth rates over all periods.However, growth rate of height showed a strong cross over during the summer in 3rd and 4th year (Fig. 2), which indicates clonal genotype × environment interaction.In the 3rd year (1997), the growth rate of height for P. tremula had one peak in the mid-June.It then decreased constantly through the rest of the growth period.We conclude that the fast growth of hybrids can be partly explained by the second growth peak in mid-August.However, the hybrid clones normally had another growth peak in mid-August, especially the height growth.In the 4th year (1998), P. tremula shows a slight tendency to form a second peak, but the hybrids behave far more erratically, responding strongly to a period of warm weather in late summer.This indicates that the hybrid genomes are adapted to two different parental environments.The late fl ushing and   early cessation of the local aspen is considered to be an adaptation to the short summers in the far north.
The signifi cant clone-by-block interaction for growth performance illustrated how sensitive aspen clones are to microsite differences (Ceulemans et.al 1992).By the end of the 5th year, block 5 in the trial produced only 48% of what block 1 produced.This must have been due to a considerable variation in soil conditions across the trial.Li and Wu (1997) found signifi cant genotype × environment interaction for growth traits of inter and intra-specifi c aspen crosses in two contrasting environments.
The ranking of clones for stem volume was not consistent from the 2nd to the 5th year.Signifi cant clone by year interaction in growth performance probably indicates differences in hybrid reaction to the climate (Table 2).Zsuffa et al. (1993) found that the ranking of poplar clones (P. × euramericana) by height changed considerably during the fi rst 6 years of growth.Our study showed similar results.Stem volume of clones 3 and 4 were the same in the 2nd year, 0.8 dm 3 .Because basal diameter of clone 3 increased faster than clone 4, the stem volume of clone 3 was larger than that of clone 4 in the 5th year.In the 2nd year, the stem volume of clones 1 and 2 were 1.1 dm 3 and 1.0 dm 3 , respectively, but after the 5th year, the stem volume were 17.6 and 17.8 dm 3 , respectively.This suggests that, for proper clonal selection several cumulative growth periods must be observed, perhaps over as long as 10 years.
The range in growth initiation was usually smaller than it was for growth cessation.The bud emerged of the hybrids started 8.85 days earlier than for P. tremula in the spring, whereas half leaves yellowed of the hybrids occurred 31.63 days later in the autumn.Compared to growth initiation, the environmental control of growth cessation is more site-specifi c.This may be due to the interaction of edaphic and weather conditions in causing rust damage.The leaf decoloration may have been indirectly affected by cold and wet weather in the 4th year.P. tremula and clone 1 had suffered from a strong infection of leaf rust, which may have affected the timing of defoliation and growth cessation.
There was a strong correlation between growth period and yield growth, which indicates that the fast growth for the hybrid can be largely explained by the expanding growth period.However, when P. tremula was removed from the data, there was no correlation between growth period and yield growth in the hybrid material (r = -0.002).Yield is a complex character, and may well be determined by other factors than the growth period.Li et al (1998) noted that hybrid vigor in aspen hybrids (P.tremula × P. tremuloides) was associated with delayed bud set resulting in longer duration of growth in fi rst year.However, by the second year, no relationship was found between bud set and stem growth in either the intra-or inter-specifi c crosses.
Strong selection for growth may increase the growth period (Wang and Tigerstedt 1996).Late active growth may render the hybrids liable to frost damage in fall and winter (Rehfeldt 1979, and Li and Adams 1993, Li et al. 1998).Li (1998) proposes the use of local aspen as female parents in intraspecifi c crosses, in order to improve frost hardiness of the hybrids.
The fast overall growth of the aspen hybrids is largely explained by their longer vegetative period.This response is not true heterosis as has been commonly assumed in hybrid aspen.It must be borne in mind that an extension of the vegetative period of about 40 days in the hybrids may drastically increase the risk of frost damage at both ends of the growth period.
The considerably extended growth period may also upset a "natural" balance between the aspen and its pests and diseases.Thus, the increased yield is likely to involve an increased risk.This situation should be carefully assessed before large-scale cultivation of cloned hybrids can take place.
Our analyses were based on a small number of clones measured over only fi ve years.The results of the study need to be followed up by larger and longer experiments.The superiority of hybrids over their non-hybrid counterparts, and the underlying cause for such superiority are unclear (Stettler et al. 1996).Valid estimates require a suffi cient number of parents and their progenies to be tested in suffi cient numbers of trials, replicated in space and time (Stettler et al. 1996).It remains to be found out, if on top of the extended vegetation period, there is in addition a sign of true heterosis that could be detected by molecular markers for heterozygosity (Aravanopoulos and Zsuffa 1998).

Fig. 1 .
Fig. 1.Mean total tree height, basal diameter, and total stem volume at the end of each of the four growing seasons for aspen hybrid clones and P. tremula.

Fig. 2 .
Fig. 2. Growth pattern of diameter at breast height (DBH), basal diameter (BD) and height (HT) in the 3rd (Figs.a,b,c) and 4th year (Figs.d,e,f) for the hybrid clones and P. tremula.For key to symbols, see Fig. 1.

Table 2 .
Mean height, basal diameter, breast height diameter and stem volume of four hybrid aspen clones and one local P. tremula seedling source at the end of 2nd and 5th year of growth.
Note: Means in vertical sequence not followed by the same letter are signifi cantly different at P < 0.05 according to Tukey' s test.

Table 3 .
Mean annual growth rates, increment of height, basal diameter and breast height diameter at different observational dates among hybrid clones and

Table 4 .
Mean values and coeffi cient of variation of the traits among the hybrid clones and P. tremula seedling source.

Table 5 .
Correlation coeffi cients (r) among the phenological and growth traits in the 3rd year.