New Forests 23: 105–119, 2002.  2002 Kluwer Academic Publishers. Printed in the Netherlands.

Changes through time in traits of poplar clones in selection trials ADRIAN ARES 1,2 1

´ , Universidad Nacional del Sur, Altos del Palihue, 8000 Bahıa ´ Blanca, Departamento de Agronomıa Buenos Aires, Argentina; 2 Current address: Dale Bumpers Research Center, Agriculture Research Service, United States Department of Agriculture, 6883 South State Highway 23, Booneville, AR 72927 -9214, USA Received 1 March 2000; accepted in revised form 13 September 2001

Key words: Age-age correlations, Argentina, Multitrait selection, Poplar silviculture, Populus deltoides Abstract. Growth, survival, stem characteristics, resistance to wind, and canker disease of 51 poplar clones were evaluated at age 9 and 10 years, and compared to measurements at age 3 in two trials on the Lower Valley of the Colorado River, Argentina. About 30% of the clones in both trials had large changes in growth rankings at age 9 and 10 compared to rankings at age 3. Canker incidence increased about 4-fold between age 3, and measurements at age 9 and 10. Stem form was in general satisfactory, and wind incidence was almost nil. In trial 1, clones 588–1, Conti 12, 582–41, and 582–5 occupied the upper positions in the growth rankings at age 3 and 10, and the clone 588–1 ranked first both in growth and canker incidence in 1998. Stem growth trajectories of clones with similar growth rates did not differ among clones with, and without canker disease. In trial 2, the range in growth among clones was narrower than in trial 1, and canker disease, although of minor incidence, was present in most clones at age 9. Some of the P. deltoides clones in the trials had similar growth, and less canker disease than the Euroamerican clones commonly grown in the region, but progressive canker incidence should be addressed with additional selection trials. ´ Resumen. Crecimiento, sobrevivencia, caracterısticas del tronco, resistencia al viento e incidencia de ˜ de edad, y comparados con ´ cancrosis fueron evaluados en 51 clones de alamos a los nueve y diez anos ˜ en dos ensayos en el Valle Inferior del Rıo ´ Colorado en el sur de la Provincia mediciones a los tres anos de Buenos Aires, Argentina. Aproximadamente 30% de los clones en ambos ensayos experimentaron ˜ en comparacion ´ con rankings a grandes cambios en los rankings de crecimiento a los nueve y diez anos ˜ los tres anos. La incidencia de cancrosis se incremento´ cerca de cuatro veces entre mediciones a los tres ˜ y a los nueve y diez anos. ˜ anos La forma del tronco fue en general satisfactoria en ambos ensayos y la incidencia del viento fue casi nula. En el ensayo 1, los clones 868-1, Conti 12, 562-41 y 562-5 ocuparon ˜ las primeras posiciones en los rankings de crecimiento a los tres y diez anos, y el clon 568-1 ocupo´ el primer lugar en crecimiento e incidencia de cancrosis en 1998. Las trayectorias en incremento diametral del tallo de clones con similar tasa de crecimiento no difirieron entre clones con y sin cancrosis. En el ´ reducido que en el ensayo 1, y sıntomas ´ ensayo 2, el rango en crecimiento entre clones fue mas de ´ de los clones. Algunos de los cancrosis, aunque de menor incidencia, se manifestaron en la mayorıa clones de P. deltoides en los ensayos tuvieron similar crecimiento y menos cancrosis que clones ´ aunque el progresivo crecimiento en la incidencia euroamericanos comunmente cultivados en la region ´ adicionales. de cancrosis requiere ensayos de seleccion

Introduction The ancient culture of poplars still draws considerable interest worldwide as a

106 source of timber, pulp, fiber, and energy. Poplars account for half of the annual wood production in Italy, and France exports some 450 000 m 3 of poplar wood per year (Steenackers et al. 1993). The area cultivated with hybrid poplars is expected to grow in countries such as Canada, Sweden, the United Kingdom, and the United States (Mitchell 1994; Heilman 1999). Poplars can contribute to slow but continuous improvements in soil conditions in forestry and agroforestry systems by increasing organic matter (Jug et al. 1999), abundance of soil fauna (Park et al. 1994), and microbiological activity (Browaldh 1995). Potentially adverse effects from poplar systems such as high water consumption may be reduced by using clones with high water-use efficiency (Allen et al. 1999) or increased drought resistance (Tschaplinski et al. 1998), while nitrate leaching and groundwater contamination could be prevented by nutrient management strategies targeted to specific sites (Jug et al. 1999). In Argentina, there are some 100 000 ha of poplar plantations. Poplars are widely cultivated on the delta of the Parana River, and on irrigated areas throughout the country. Among these areas, the Lower Valley of the Colorado River (LVCR) in Southern Buenos Aires comprises 700 000 ha of which about 90 000 ha are irrigated for crop production and pastures. More than 70% of the poplar plantations in the region are constituted by the Euroamerican clones P. x canadensis cv. I-214 and Conti 12 which replaced P. nigra L. cv italica, the first introduction in the region (Gutierrez and Ares 1996). All these clones became susceptible to canker disease: therefore, tree health is of primary concern in selecting poplar clones for the region. Canker disease in the LVCR is supposedly caused by the fungus Septoria musiva Peck., but no isolations of the pathogen have been made so far, and other organisms may be involved. Most of the poplar wood harvested in the LVCR is used for fruit boxes, pallets, and economical housing, but there is potential for other products if better quality wood is produced. Traits related to tree form are of interest because of the current use of rotary cutting, and the potential of processing wood for plywood and veneer. For that purpose, straight and non-forked stem, flat branching habit, and small branches are preferred tree characteristics. Previous research in Argentina (Scordo and Arreghini 1983) and elsewhere (Newman 1997; Scarascia-Mugnozza et al. 1997) has indicated that Populus deltoides Bartr. and several of its hybrids often have rapid growth, resistance to canker disease, high survival, and good stem characteristics. Therefore, P. deltoides clones may provide an alternative to Euroamerican clones in the LVCR. Trials on representative sites should be carried out long enough to provide reliable assessment of clone performance. It has been suggested that final selection should not be made earlier than at one half the rotation age (Zobel and Talbert 1984), or even at one entire rotation period (Lo et al. 1995). Long-term trials also allow retrospective testing by examining in vitro, or during early growth stages, the response of clones which have already been tested under field conditions. In that way, it would be possible to validate growth and disease resistance responses, and examine underlying mechanisms and processes. In this study, several characteristics of 51 poplar clones were evaluated at ages 9

107 and 10 years in two field trials, and compared to measurements at age 3. Traits related to tree growth, survival, stem form, wind and disease resistance were used to assess clone performance. It was hypothesized that indices measured at 3 years of age predicted future performance of poplar clones, and that canker incidence increased through time and was detrimental to tree growth.

Materials and methods The study area The experimental site was on an intermediate alluvial terrace on the LVCR, southern Buenos Aires, Argentinay (Figure 1). The soil was a loamy sand Typic Fluvaquent (Soil Survey Staff 1992) with weak structure and organic matter content of 1.5%. The water table was about 1.8 m deep. Mean annual rainfall is 368 mm, and water deficits may occur in all months except May, July, and October when precipitation exceeds mean evapotranspiration. Mean annual temperature is 13.8 8C with monthly means of 22.2 8C in January, and of 7 8C in July. Mean annual frequency of frozen days is 92, and strong winds, mainly from the N, NW and W are common. The area to be planted was prepared for irrigation in 1987, and barley and alfalfa

Figure 1. Location of the Lower Valley of the Colorado River and other areas with poplar cultivation in Argentina.

108 were sown. Before the start of the trials, the area was ploughed and disked. None of the trials were fertilized. Trial 1 In September 1988, 40-cm long rooted cuttings of 26 poplar clones were planted by hand at a spacing of 4 m by 4 m. A randomized complete block design with one plant per plot and six replicates was set up. The plantation was surrounded by a border row of the P. deltoides clone 588-1. Irrigation was carried out twice a week at a rate of 1350 mm ha 21 yr 21 . Failed plants were replaced with one-year old plants within the year following the original planting. Weeds were controlled with five mechanical weedings during the first year, and three mechanical weedings per year in the subsequent years until the end of the third year. Manual weedings were two per year until the third year. Twenty-two of the clones were selections of P. deltoides obtained by controlled fertilization by the National Institute of Agricultural Technology of Argentina (INTA) about 700 km north of the testing site. The P. deltoides clones Australia 129–60 and Harvard (formerly 63–51), and the Euroamerican clones Conti 12 and I-488 were also included in the trial, the latter two as controls. Trial 2 In August 1989, 40-cm long rooted cuttings of 28 poplar clones were planted by hand at a spacing of 3 m by 3 m. Square plots with four trees each were laid out in a randomized complete block design with six replicates. The plantation was surrounded by a guard row of the clone Catfish 5. Irrigation and plant replacement were carried out at as in trial 1. Weeds were controlled with six mechanical and two manual weedings during the first year, and three mechanical weedings per year in the subsequent years until the end of the third year. Twenty-five out of 47 clones of P. deltoides were selected in the nursery. Sixteen clones were from selections made in Argentina by INTA, and nine clones were from Australia. Italy, and the United States. Four clones were hybrid. The clone Conti 12 was used as control. Field measurements and statistical analysis Total tree height was recorded in 1991 and 1998 in trial 1, and in 1990, 1992 and 1998 in trial 2. Height measurements were done with a wooden rule initially, and with an altimeter later. Stem diameter outside bark at 1.37 m above ground (DBH) of all trees in the trials was calculated from tree girth measured in 1991, 1994, 1996 and 1998 in trial 1, and in 1992 and 1998 in trial 2. A scaled volume calculated as DBH 2 times total height was used as a growth index as in other tree selection studies (Rogers et al. 1988; Abrahamson et al. 1990). Survival was evaluated during the second year in both trials and the measurements included the replacement trees. In 1991 on trial 1, and in 1992 on trial 2, the trees were visually assessed for stem

109 and branch canker, leaf rust, wind damage, branch characteristics, and stem form as previously detailed (Ares and Gutierrez 1996). Briefly, the canker severity index ranged between 0 5 no cankers to 3 5 four or more cankers, or more than half of the stem girdled; rust severity was between 0 5 no infection to 3 5 . 50% of the leaf affected; wind damage was between 0 5 no damage to 2 5 large branches and / or the leading top broken; branch diameter at the point of attachment to the bole was 0 5 , 10 mm to 3 5 . 30 mm; forking was 0 5 no forks to 2 5 multiple forks, and crookedness was 0 5 straight stem to 2 5 stem with two crooks or more. Branch angle was measured with a protractor on five randomly selected branches in each tree, and taper was calculated as tree height divided by DBH. This ratio is a rough estimate of taper as trees with the same height and DBH can have different form. Rust severity was recorded before leaf abscission in 1991 and 1992, but it was not recorded in 1998. Branch angle was not measured in 1998 as previous data showed that this trait had a narrow range of variation and no statistical differences among clones. Comparisons of means for the different traits among clones were performed with the Tukey’s studentized range test for unequal number of cases. Pearson age-age correlation analysis was used to measure the association between trait values at different ages. In trial 1, continuous and categorical variables for the four hybrid clones were compared to those for P. deltoides clones by using t-test and ordinal logistic regression. Clones were grouped by non-hierarchical (k-means) cluster analysis (Hartigan 1975) on standardized data including mean values for each trait. All traits were given equal weight in the analysis. Discriminant analysis was used to determine by a jacknife cross validation the percent of clones correctly assigned to the groups obtained by cluster analysis. The purpose of both the cluster and discriminant analyses was to infer which groups of clones had the highest and lowest probability of success for cultivation on areas of the LVCR similar to the testing site. Analysis of variance of repeated measures was performed on stem diameter data measured in 1988, 1991, 1996 and 1999 in trial 1. The coefficients for the single degree of freedom tests were adjusted for the uneven spacing between measurements to partition the sums of squares differently. The Wilk Shapiro Rankit Plot test and the Bartlett’s test of equal variances were used to test assumptions of the ANOVA analysis. All statistical analyses were performed with the SYSTAT 7.0 statistical software (SYSTAT 1997). An alpha 5 0.05 was used in all comparisons to determine significant differences.

Results Trial 1 Average annual DBH increment ranged between 0.46 and 2.77 cm, and average annual height increment varied from 0.57 to 1.54 m among clones. Average DBH and total height at age 3 were significantly correlated to average DBH and total height at age 10 (r 5 0.87, P , 0.01 for DBH, and r 5 0.90, P , 0.01 for total

110 height). The clone 568-1 ranked first both in scaled volume and canker index (Table 1). There were no significant differences in scaled volume, however, among the top seven clones including the Euroamerican clone Conti 12. The clones 568-1, Conti 12, 562-41, and 562-5 occupied the upper four positions for scaled volume at age 3 and at age 10. Nonetheless, some clones changed considerably in their rank position between age 3 and age 10. The clone 610–12 ranked fourteenth in scaled volume at age 3 and fifth at age 10, while the clone 564–10 dropped from sixth at age 3 to nineteenth at age 10. Despite changes for some clones, scaled volume at age 3 was positively correlated to scaled volume at age 10 (r 5 0.94, P , 0.01). Seventeen out of 26 clones had no visible evidence of canker disease at age 3, whereas 11 clones out of 26 clones had no evident canker disease at age 10 (Table 1). The clone 568-1 was apparently free of canker disease at age 3, but it turned to be severely affected at age 10. Three clones (562-41, 610-12 and 610-16) among the seven clones with high scaled volume at age 10 (Table 1) had no canker disease. Table 1. Poplar clones in trial 1 ranked by scaled volume (SVOL; cm 2 m) and canker incidence (CAN). Both at age 3 and at age 10, survival of all clones was 100% except for clone 565-4 (75%). Comparisons of means by Tukey’s studentized range test are included for SVOL at age 10. Means for SVOL are followed by one standard error. Values followed by the same letter do not differ at a 5 0.05. All clones are P. deltoides except the clones Conti 12 and I-488 (P. x canadensis) which are noted with an asterisk. Age 10

Age 3

Clone

SVOL

CAN

Rank

SVOL

Rank

CAN

Rank

568-1 Conti 12 * 562-41 562-5 610-12 610-16 129-60 I-488 * 610-24 610-1 582-1 564-17 610-11 608-15 562-4 582-4 564-13 Harvard 564-10 565-9 565-7 568-3 565-4 565-8 580-6 565-3

12187 6 2004 a 8750 6 2024 ab 8632 6 1522 ab 8235 6 2565 ab 5727 6 2911 abc 5705 6 1671 abc 5598 6 1615 5365 6 508 bc 5207 6 1391 bc 5068 6 1296 bc 4770 6 1428 bc 4769 6 1382 bc 4187 6 963 bc 3804 6 1362 bc 3684 6 1022 bc 3564 6 1572 3525 6 897 bc 3467 6 1081 bc 3094 6 2181 bc 2937 6 1207 bc 2235 6 1059 bc 2184 6 1117 bc 1326 6 609 c 1191 6 582 c 795 6 533 c 299 6 235 c

3.00 1.25 0.00 0.25 0.00 0.00 0.75 2.75 0.25 0.75 1.50 0.00 0.00 0.00 0.00 0.25 0.00 0.00 0.00 0.50 0.25 1.25 0.25 0.25 0.25 0.00

1 4 15 9 15 15 6 2 9 6 3 15 15 15 15 9 15 15 15 8 9 4 15 9 9 15

577 348 381 330 174 196 213 215 199 225 239 106 231 148 181 172 103 158 234 125 145 57 52 24 26 11

1 3 2 4 14 12 10 9 11 8 5 20 7 17 13 15 21 16 6 19 18 22 23 25 24 26

0.00 0.50 0.00 0.00 0.00 0.00 0.25 1.00 0.25 0.25 0.50 0.00 0.00 0.0 0.00 0.00 0.00 0.00 0.00 0.00 0.25 0.50 0.00 0.25 0.00 0.00

10 2 10 10 10 9 5 1 5 5 2 10 10 10 10 10 10 10 10 10 5 2 10 5 10 10

111 Canker incidence at age 3 was correlated to canker incidence at age 10 (r 5 0.67, P , 0.01). Average survival in trial 1 was 90%. At age 10, forking was minor ranging between 0 and 0.5, while crookedness varied between 0.25 to 1.25 among clones. For both variables, there were no significant differences among clones (P 5 0.42 for forking, and P 5 0.32 for crookedness). The index for branch diameter ranged between 0 and 1.5, and taper varied between 0.57 and 1.20 with significant differences among clones for both variables (P , 0.01 for branch diameter, and P , 0.01 for taper). Wind damage was almost nil for all clones. Indices related to stem form measured at age 3 and at age 10 were correlated (r 5 0.60, P , 0.01 for crookedness; r 5 0.51, P 5 0.01 for branch diameter, and r 5 0.58, P 5 0.02 for forking). The four hybrid P. deltoides clones did not differ to the P. deltoides clones for most traits except canker incidence (P 5 0.01) which, in average, was higher in the hybrid clones. Five groups of clones were distinguished by cluster analysis based on all recorded traits except survival (Table 2). Group 1 included clones with low growth and low canker disease, and relatively good stem characteristics such as nil fork, few large branches, and fairly straight stems. Clones in group 2 had moderate growth, low canker disease, good stern form, and moderate incidence of large branches. Group 3 comprised clones with moderate growth and low canker disease, and moderate number of large branches and crooked stems. Clones in group 4 had high growth rates but moderate to high canker disease, moderately high incidence of large branches, and some incidence of forkiness and crookedness. Group 5 contained clones characterized by distinctively high growth and canker disease, moderately high number of large branches, and moderate crookedness. In general, clones with high growth had lower taper and more large branches. On average, 96% of the clones were correctly reclassified within the groups obtained by cluster analysis. Clustering of clones based only on scaled volume and canker index produced four groups (Figure 2a). Clones in groups 1 and 2 were separated by their growth rates as the degree of canker incidence was similarly low. Group 3 included clones with moderate canker disease and varied growth rates. Clones in group 4 had moderate to high growth but high canker incidence. Sequential measurements of DBH in trial 1 allowed the examination of stem growth trajectories through time. The DBH values were normally distributed (WilkTable 2. Composition and average characteristics of groups of clones in trial 1 obtained by k-means clustering at age 10. Grouping was based on: scaled volume (SVOL, cm 2 m), canker incidence (CAN), forking (FOR), branch diameter (BD), crookedness (CRO) and taper (TAP; m cm 21 ). Allowed ranges were 0–3 for CAN, 0–2 for FOR, 0–3 for BD and 0–2 for CRO. Group Clones

SVOL CAN FOR BD CRO TAP

1 2 3 4 5

1490 4294 4041 6399 9856

564-10, 565-3, 565-4, 565-7, 565-8, 582-6 562-4, 564-13, 564-17, 568-3, 610-1, 610-11, 610-12, 610-24 Harvard, 565-9, 582-1, 582-4, 608-15, 610-16 129-60, 1-488, 562-5 Conti 12, 562-41, 568-1

0.17 0.28 0.37 1.25 1.42

0.00 0.03 0.08 0.42 0.00

0.21 0.50 0.75 1.33 1.42

0.67 0.50 1.08 0.58 0.75

1.07 0.77 0.78 0.69 0.60

112

Figure 2. Groups of clones (identified as C) in (a) trial 1, and (b) trial 2 obtained by means clustering analysis based on scaled volume (SVOL) and canker disease (CAN) indices. Vertical lines indicate separation among groups not actual cluster boundaries.

Shapiro index 5 0.99) and residuals had equal variances (P 5 0.96 for Bartlett’s test). For the four groups obtained by cluster analysis based on scaled volume and canker index, there were significant linear, quadratic and cubic effects of time on the average trajectories. The linear trend, however, accounted for 98% of the variability in DBH through time (P , 0.01). Changes through time of the trajectories were not the same for all groups as indicated by the significant interaction between time and group (P , 0.01). The trajectories of groups 2, 3 and 4 did not differ (Figure 3)

113 although probability levels were close to significance (P 5 0.09 for the contrast of group 2 versus group 3; and P 5 0.08 for the contrast of group 3 versus group 4). Trajectory of clones in group 1 having low growth and very low canker incidence differed from the trajectories of all other groups (P , 0.01). The similarity between trajectories of group 2 including clones with moderate growth and low canker incidence, and group 4 (clones with both moderate growth and canker incidence) suggested that canker disease did not affect patterns of growth through time (Figure 3). Accordingly, the interaction canker index by clone was not significant for growth rates of clones in Trial 1 (P 5 0.28). As examples, the trajectory of clone 568-1 which had high growth rate and canker disease, did not differ to that for clone 562-41 which had also high growth rate, but no canker incidence. Also, trajectory of clone 568-3 having low growth and high canker disease was similar to that for clone 564-10 which had low growth and no canker diseease. Trial 2 Average annual DBH increment ranged between 1.53 to 2.28 cm and, average annual height increment varied from 1.32 to 2.22 m among clones. Average DBH and total height at age 9 were correlated to average DBH and total height at age 3 (r 5 0.76, P , 0.01, and r 5 0.44, P 5 0.03 for total height). Also, average total height at age 1 was correlated to average total height at age 3 (r 5 0.72, P , 0.01),

Figure 3. Growth in stem diameter of P. deltoides clones characterized by: (i) low growth and low canker incidence (Group 1), (ii) moderate growth and low canker incidence (Group 2), (iii) high growth and severe incidence of canker disease (Group 3), and (iv) moderate growth and moderate canker incidence (Group 4), in trial 1.

114 and at age 10 (r 5 0.45, P 5 0.02). The range in scaled volume at age 9 was narrower than that for trial 1, with about a 2.4-fold difference between the clones with the highest and the lowest growth rates (Table 3). As in trial 1, the clone Conti 12 ranked second in scaled volume. Because of large variability among individuals of each clone, however, there was no significant differences in scaled volume among most clones at age 9. The relative position of several clones along the rank in scaled volume changed considerably between age 3 and age 9. The clone Stoneville 72 ranked tenth in scaled volume at age 3 and first at age 9, while the hybrid clone 186-68 dropped from first place at age 3 to sixteenth at age 9. Scaled volume at age 3 was positively correlated to scaled volume at age 9 (r 5 0.63, P , 0.01). Incidence of canker disease became more prevalent between age 3 and age 9. At age 3, twenty clones had no visible evidence of canker incidence compared to a single clone (Stoneville 71) without canker disease at age 9 (Table 3). Most clones Table 3. Poplar clones in trial 2 ranked by scaled volume (SVOL cm 2 m 21 ), survival (SUR; %) and canker incidence (CAN). Comparison of means by Tukey’s studentized range test are included for SVOL at age 9. Means for SVOL are followed by one standard error. Values followed by the same letter do not differ at a 5 0.05. Clones followed by an asterisk are hybrids; the rest are P. deltoides except the clone Conti 12 (P. x canadensis). Age 9

Age 3

Clone

SVOL

SUR Rank CAN Rank SVOL Rank SUR Rank CAN Rank

Stonville 72 Conti 12 208-68 * Fierolo 129-60 Spiado 217-68 * Stonville 109 71-67 Catfish 5 10-67 233-69 106-60 347-69 Stonville 107 186-68 * Stonville 71 341-69 107-68 * Catfish 2 346-69 Stonville 92 167-69 350-69 369-69 37-67

74486524 a 70086713 a 680462345 ab 66566420 ab 662761154 ab 60726754 ab 60606715 ab 60396436 ab 59186688 ab 57736762 ab 53676377 ab 53606808 ab 53596611 ab 52366522 ab 51406553 ab 50376522 ab 483261186 ab 37766360 ab 37336841 ab 36486664 ab 36286495 ab 35626483 ab 33996617 ab 33516508 ab 33256455 ab 31386467 b

50 96 87 100 79 100 100 83 96 100 96 100 92 92 100 100 38 100 75 96 100 96 100 100 87 100

25 13 20 1 23 1 1 22 13 1 13 1 18 18 1 1 26 1 24 13 1 13 1 1 20 1

0.33 1.74 0.13 0.42 0.26 0.79 0.67 0.17 0.74 0.58 0.52 0.50 0.18 0.45 0.25 1.29 0.20 0.42 0.94 0.09 0.29 0.39 0.42 0.46 0.19 0.29

16 1 25 12 19 4 6 24 5 7 8 9 23 11 20 2 21 12 3 26 17 15 12 10 22 17

547 712 372 657 586 639 560 518 674 501 640 496 489 679 439 764 310 374 466 250 379 318 455 357 384 379

10 2 22 5 8 7 9 11 4 12 6 13 14 3 17 1 25 21 15 26 19 24 16 23 18 19

50 96 87 100 79 100 100 83 100 100 100 100 92 100 100 100 38 100 75 96 100 96 100 100 92 100

25 16 21 1 23 1 1 22 1 1 1 1 19 1 1 1 26 1 24 16 1 16 1 1 19 1

0.00 2.04 0.00 0.00 0.00 0.00 0.00 0.00 0.67 0.00 0.00 0.00 0.00 0.00 0.00 0.08 0.00 0.33 0.00 0.17 0.00 0.00 0.00 0.00 0.00 0.21

7 1 7 7 7 7 7 7 2 7 7 7 7 7 7 6 7 3 7 5 7 7 7 7 7 4

115 at age 9, however, had relatively minor incidence of canker disease. Canker incidence at age 3 was correlated to canker incidence at age 9 (r 5 0.67, P , 0.01). Average survival in trial 2 was 91%. Survival at age 3 was highly correlated to survival at age 9 (r 5 0.99, P , 0.01). In 1998, forking ranged between 0 and 0.61, while crookedness varied between 0.44 and 1.50 among clones. There were significant differences among clones for both forking (P , 0.01) and crookedness (P , 0.01). The range for the branch diameter index was 0.48 to 1.83, and it was 0.78 to 1.16 for taper. There were significant differences among clones for branch diameter (P , 0.01), but there were no differences for taper (P 5 0.21). As in trial 1, wind incidence was almost nil. Most indices related to stem form measured at age 3 and age 9 were correlated (r 5 0.67, P , 0.01 for crookedness; r 5 0.46, P 5 0.02 for branch diameter, and r 5 0.40, P 5 0.04 for taper). Five groups of clones were suggested by cluster analysis based on all recorded traits except wind damage (Table 4). Group 1 included clones with low growth and high survival. Clones in group 2 had high growth and survival, but some undesirable stem characteristics such as relatively high incidence of crookedness and large branches. Clones in group 3 had high growth, low survival, and fairly good stem characteristics. Group 4 comprised clones with moderate growth, moderate survival, and stem characteristics similar to clones in group 3. Clones in group 5 had moderate to high growth, high survival, and relatively poor stem characteristics. Canker incidence was similar among groups of clones. In average, 88% of the clones were correctly reclassified within the groups obtained by cluster analysis. Clustering of clones in trial 2 based only on scaled volume and canker determined five groups (Figure 2b). Group 1 included clones with both low growth and canker disease. Separation between groups 2 and 3 was mostly based on canker incidence (0.24 in average for group 2, and 0.61 for group 3). The clone Conti 12 with both high growth and canker incidence constituted a one-clone group, while two clones with both moderate canker incidence and growth were included in group 5. On average, 92% of the clones were correctly reclassified within groups.

Table 4. Composition and average characteristics of groups of clones in trial 2 obtained by k-means clustering at age 9. Grouping was based on: scaled volume (SVOL, cm 2 m), canker incidence (CAN, survival (SUR, %), forking (FOR), branch diameter (BD), crookedness (CRO) and taper (TAP; m cm 21 ). Allowed ranges were 0–3 for CAN, 0–2 for FOR, 0–3 for BD and 0–2 for CRO. Group

Clones

SVOL

CAN

SUR

FOR

BD

CRO

TAP

1

Catfish 2, Stoneville 92, Stoneville 107, 37-67, 167-69, 341-69, 346-69, 350-69, 369-69 Fierolo, Spiado, 10-67, 71-67 Stoneville 71, Stoneville 72, Stoneville 109, 129-60 248-68, 107-68 Catfish 5, Conti 12, 106-60, 188-68, 217-68, 233-69, 347-69

3663

0.31

97.7

0.10

0.86

0.88

0.91

6003 6236

0.62 0.24

98.0 62.5

0.41 0.10

1.56 0.85

1.26 0.59

0.84 0.86

5268 5694

0.53 0.77

81.0 97.1

0.03 0.15

0.78 0.90

0.89 0.70

1.07 0.81

2 3 4 5

116 Discussion Several poplar clones in the selection trials maintained similar features of growth, health, and stem form through time. In trial 1, the same clones occupied the upper four positions in the growth ranking at age 3 and at age 10. Characteristics of other clones, however, changed greatly as the trials progressed. About 30% of the clones in trials 1 and 2 had fairly large changes in their relative position in the rank in scaled volume as noted previously for other growth traits such as height (Zsuffa 1975). The clone Stoneville 72 in trial 2 was within a second-rank group at age 3, and rose to first position in scaled volume at age 9. Its performance at age 9 cannot be attributable to release of competition in the trial plots because of rapid growth even in plots without tree mortality. For this clone, however, low survival may be a silvicultural problem if missing plants cannot be successfully replaced initially. In general, low survival of some clones was likely related to rooting characteristics during early growth stages because tree mortality did not increase after three years in most clones. Overall, survival did not likely modify trial results to a large extent. In both trials, survival was not correlated to growth or canker resistance. In general, DBH and total height measured at ages 9 and 10 years were correlated with data at age 3. In trial 2, total height at one year correlated to total height at age 3, but the correlation was weaker between ages 1 and 9, and between ages 3 and 9. Differences among clones in sensitivity to frost damage, biomass allocation, and branching patterns would be possible explanations to differences in age-age correlations between trials. Results from these trials were somewhat different to some previous studies with hybrid poplars showing that height at early stages did not predict height at later stages in the rotation (Hansen et al. 1992; Brown et al. 1996), In a provenance trial with 40 clones of P. deltoides, however, there were large age-age correlations for height, DBH and stem volume (Foster 1988). Growth rates in the trials were moderate and similar to previous measurements on stands of P. x euroamericana and P. alba clones in the region (Ares and Gutierrez 1993), although the reduced size of the trial plots did not allow us to determine wood production. Survival after replacement for failures one year after planting was generally high, and if that is also the case in larger stands, survival will not exert the strong effect on volume either per tree or per unit area that can occur with lower survival rates (Chambers and Borralho 1997). There was increasing incidence of canker disease through time with about a four-fold increase in average between age 3, and ages 9 or 10 in both trials. Increased canker incidence through time agreed with the results of a poplar trial in northern New York State in which canker disease increased on about 28% of the clones over a six-year period (Lo et al. 1995). This progressive advance of canker disease may occur in a similar way in large plantations in the LVCR. In some, albeit exceptional cases, the change in canker incidence was quite remarkable such as the clone 568-1, which was apparently free of canker disease at age 3, and turned to be severely affected at age 10. Indeed, the extent of canker incidence on every individual of clone 568-1 in the trial was the maximum allowed in the proposed scale. Similar growth trajectories observed in

117 trial 1 on clones with and without canker incidence suggested that stem growth rates were not reduced by the disease. Growth rates, however, may have been higher if the clones had been free of canker disease. In general, the tested P. deltoides clones had acceptable stem characteristics in the two spacings at 3 by 3 m, and 4 by 4 m. There was low incidence of forkiness and moderate crookedness, despite slightly waved stems were counted as crooked. Stem characteristics such as branch diameter and taper can change with stand age and density (Steenackers et al. 1993; Muhairwe 1994; DeBell and Gartner 1997), and also with thinning, pruning, and fertilization. Among seven clones in trial 1 recommended in 1992 for larger trials across the region, the clones 129-60, 562-5 and 562-41 remained within the group with high growth rate, and had nil to low canker incidence. Interestingly, the clone 129-60 had similar characteristics and rank, positions in both trials. The clone 562-41 was especially valuable with no canker incidence, no forkiness, and moderate crookedness. Clones such as 564-17, 610-12, 610-16, and 610-24 included in group 2 based on scaled volume and canker could be also recommended for further study. The clone 610-12 had desirable stem characteristics such as nil fork, very slight crookedness and few large branches. The clones in trial 2 recommended for larger trials based on their early performance (Ares and Gutierrez 1996) had varied characteristics in 1998. The clone Fierolo continued to rank high in growth and survival, but had some canker disease. The hybrid clone 186-68 considered the most promising clone at age 3 was the second most affected by canker disease at age 9. The clones 71-67 and 347-69 had moderate growth (although not significantly different from most clones) and some canker incidence at age 9. Measurements at age 9 indicated that several clones in groups 2, 3, 4, and 5 would be suitable for future trials. Exceptions are clones such as Stoneville 71 with only 38% survival, and 45% less growth than the top clone. Canker incidence, however, should be a matter of concern as canker incidence may increase with time. The hybrid clone 208-68 is probably the most valuable in the trial because it has high growth, very low canker incidence (12% of the individuals minimally affected), high survival, and acceptable stem form. No clone showed the best scores for all characteristics. Large variability among individuals of the same clone reduced the possibility of providing strong recommendations. Such variability is somewhat surprising for identical genetic material, and may be related to acute clonal sensitivity to microsite conditions. No measures of variability are presented in related studies where growth indices are often obtained by summing up figures for individual trees (Lo et al. 1995). A limitation of trial 1 was the low number of trees for each clone. This study indicated that growth potential, survival, stem form, and canker disease can be generally predicted from early growth stages, but the performance of some clones may change dramatically through time. The risk of providing imperfect recommendations can be reduced by working with groups of clones, rather than with individual clones. It is important to note that conclusions of this trial were drawn for

118 poplar clones growing in small plots. Some traits such as disease incidence and growth would be affected when poplar clones are planted in large pure clone stands and also under different fertilization regimes. Some of the P. deltoides clones in the trials had similar growth, less canker disease, and similar or better stem form than the Euroamerican clones commonly grown in the region. However, the progressive incidence of canker disease poses a problem in recommending clones for the LVCR. For now, an alternative would be to rely upon the P. deltoides clones with lower canker disease, and to start tests with another poplar genotypes within the different Populus sections. Overall, yield selection trials should be carried out long enough to provide a proper assessment of the tested genetic material as indicated in previous poplar selection trials (Lo et al. 1995).

Acknowledgements ´ Colorado in Pedro Luro, Argentina for I am deeply grateful to CORFO Rıo supporting this research. I would also wish to thank O. Svachka for supervising the maintenance of the trials, L. Gutierrez for starting the trials in 1988 and 1989, G. Arias for valuable field assistance, and J. Deenik, R. Arias and two anonymous reviewers for valuable comments on early drafts of the manuscript.

References Abrahamson L.P., White K.H., Nowak C.A., Briggs R.D. and Robinson D.J. 1990. Evaluating hybrid poplar clonal growth potential in a three-year-old genetic selection field trial. Biomass 21: 101–114. Allen S.J., Hall R.L. and Rosier P.T.W. 1999. Transpiration by two poplar varieties grown as coppice for biomass production. Tree Physiol 19: 493–501. ´ Colorado. Ares A. and Gutierrez L. 1993. Aspectos de la populicultura en el Valle Inferior del Rıo ´ difundidos. Rev. Asoc. For. Arg. 47: 49–60. Comportamiento de las especies y clones mas Ares A. and Gutierrez L. 1996. Selection of poplar clones for the Lower Valley of the Colorado River, Argentina. Forestry 69: 75–82. Browaldh M. 1995. The influence of trees on nitrogen dynamics in an agrisilvicultural system in Sweden. Agrofor. Syst. 30: 301–313. Brown K.R., Bead F.D. and Hogan G.D. 1996. Establishment-year height growth in hybrid poplars; relations with longer-term growth. New For. 12: 175–186. Chambers P.G.S. and Borralho N.M.G. 1997. Importance of survival in short rotation tree breeding programs. Can. J. For. Res. 27: 911–917. DeBell J.D. and Gartner B.L. 1997. Stem characteristics on the lower log of 35 year old western red cedar grown at several spacings. Western J. Appl. For. 12: 9–14. Foster G.S. 1988. Provenance variation of eastern cottonwood in the Lower Mississippi Valley. Silvae Geneticae 35: 32–38. ´ Colorado. Nivel Gutierrez L. and Ares A. 1996. Aspectos de la populicultura en el Valle Inferior del Rıo ´ tecnico de la actividad. Rev. Asoc. For. Arg. 50: 71–78. Hansen E., Heilman P.E. and Strobl S. 1992. Clonal testing and selection for field plantations. In: Mitchell C.P., Ford-Robertson J.B., Hinckley T. and Sennerby-Forese I., Ecophysiology of Short-rotation Forest Crops. Elsevier Applied Science Publishers, Oxford, UK, pp. 124–145. Hartigan J.A. 1975. Clustering Algorithms. John Wiley and Sons Inc., New York.

119 Heilman P.E. 1999. Planted forests: poplars. New For. 17: 89–93. Heilman P.E. and Xie F. 1993. Influence of nitrogen on growth and productivity of short-rotation Populus trichocarpa x Populus deltoides hybrids. Can. J. For. Res. 23: 1863–1869. Jug A., Makeschin F., Rehfuuess K.E. and Hofmann-Schielle C. 1999. Shortrotation plantations of balsam poplars, aspen and willows on former arable land in the Federal Republic of Germany. III. Soil ecological effects. For. Ecol. Manage 121: 85–99. 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. Mitchell C.P. 1994. Developments in the production and supply of wood for energy. Renewable Energy 5: 754–761. Muhairwe C.K. 1994. Tree form and taper variation over tine for interior lodgepole pine. Can. J. For. Res. 24: 1904–1913. Newman S.M. 1997. Poplar agroforestry in India. For. Ecol Manage. 90: 13–17. Park J., Newman S.M. and Cousins S.H. 1994. The effects of poplar (P. trichocarpa x deltoides) on soil biological properties in a silvoarable system. Agrofor. Syst. 25: 111–118. Rogers D.L., Stettler R.F. and Heilman P.E. 1988. Genetic variation and productivity of Populus trichocharpa and its hybrid. III. Structure and pattern of variation in a 3-year field test. Can. J. For. Res. 19: 372–377. Scarascia-Mugnozza G.E., Ceulemans R., Heilman P.E., Isebrands J.G., Stettler R.F. and Hinckley T.M. 1997. Production physiology and morphology of Populus species and their hybrids grown under short rotation. II. Biomass components and harvest index of hybrid and parental species clones. Can. J. For. Res. 27: 285–294. ´ de 19 clones de Populus sp. Scordo J.C. and Arreghini R.I. 1983. Ensayo comparativo de adaptacion ´ (Mendoza) Proc. 5 th Argentinean Forestry Congress., pp. 1122–1126. cultivados en Tunuyan Soil Survey Staff 1992. Keys to Soil Taxonomy, SMSS Technical Monograph No. 19. 5th edn. Pocahontas Press, Blacksburg, USA. 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. Chron. 69: 730–735. SYSTAT 1997. SYSTAT 7.0 Statistics. Chicago, USA. Tabbush P. and Beaton A. 1998. Hybrid poplars: present status and potential in Britain. Forestry 71: 355–364. Tschaplinski T.J., Tuskan G.A., Gebre G.M. and Todd D.E. 1998. Drought resistance of two hybrid Populus clones grown in a large-scale plantation. Tree Physiol. 18: 653–658. van den Driessche R. 1999. First-year growth response of four Populus trichocarpa x Populus deltoides clones to fertilizer placement and level. Can. J. For. Res. 29: 554–562. Zobel B. and Talbert J. 1984. Applied Forest Tree Improvement. J. Wiley and Sons, New York, USA. Zsuffa L. 1975. Some problems of hybrid poplar selection and management in Ontario. For. Chron. 51: 240–242.