Abstract: Four Populus clones were grown at three spacings (0.5 m, 1.0 m, and 1.5 m) in monoclonal and polyclonal plots in western Washington. After the third year, many individual tree and stand traits differed significantly by clone, spacing, deployment method, and their interactions. Differences among clones in growth and form were greater in polyclonal than in monoclonal plots, and differences in performance between deployment methods were greater in the denser spacings. Monoclonal stands had greater uniformity in tree size than polyclonal stands. Total woody yield decreased with increased spacing. Some clones differed in yield from other clones in both monoclonal and polyclonal plots. Assuming equal numbers of plants from the same clones were planted, the manner of deployment had no effect on productivity; that is, although there were clonal differences in yield, mean yield of the four clones in monoclonal plots did not differ from the yield of polyclonal plots. Comparative yields (polyclonal ÷ monoclonal) of the clones in polyclonal plantings differed substantially, however, and the increases or decreases in comparative yield differed with spacing. As a result, production and inventory was less evenly balanced among clones with polyclonal than monoclonal deployment.

Short-rotation intensive culture (SRIC) of clonal poplar and willow plantations has advanced from a theoretical concept to a viable fiber and biomass production system through strong research and development efforts in North America (Ranney and others 1987; Richardson 1989) and Europe (Christersson and others 1993). In the northwestern USA, research on genetics and physiology has produced several hybrid poplar clones (Stettler and others 1988; Hinckley and others 1989) that are very productive when planted on suitable sites using appropriate cultural techniques (Heilman and others 1991). SRIC is becoming an important component of the rapidly changing forest products economy of the Pacific Northwest (Miner 1990) where several companies have established large farms to produce poplar fiber.

Although current knowledge is sufficient to establish productive Populus plantations, significant questions remain concerning effects of spacing and genotype deployment on growth and yield. Most clones have been selected based on growth performance in small, monoclonal evaluation plots of a single spacing. Few data have been collected to evaluate or compare growth and yield of clones in larger plots or in plots of different spacings. One European study has shown that relative growth of clones may differ by spacing (Panetsos 1980), but experimental environment (i.e., different clones planted on adjacent spokes in a Nelder's design) consisted of inter-clonal as well as intra-clonal competition. There is little information concerning the degree to which relative clonal performance in monoclonal planting changes with spacing. Several reviews have considered factors to be considered in decisions about clonal deployment (DeBell and Harrington 1993; Lindgren 1993; Zsuffa and others 1993; Foster and Knowe 1995). There is a paucity of experimental data that address specific questions related to monoclonal vs. polyclonal deployment. Preliminary reports exist for a small test in Yugoslavia (Markovic and Herpka 1986) and one in Oregon (Shuren 1994, 1996); in addition, diameter distributions have been modeled using data from several combinations of two clone mixtures of eastern cottonwood (Populus deltoides Bartr.) (Knowe and others 1994). Do clones grow similarly (in absolute terms and relative to each other) in monoclonal and polyclonal plots? Are there differences in yield between monoclonal and polyclonal plots? Do the answers to these questions concerning deployment differ with spacing?

To help answer the above questions, we established monoclonal and polyclonal plantings of four Populus clones at three spacings. This paper reports 3-year results for survival, height, diameter, tree form, stand uniformity, and yield as affected by spacing in monoclonal and polyclonal plots.

Materials and Methods

Site description

The research plantings for our study were established in spring 1990 at the Washington State Department of Natural Resources Meridian Seed Orchard, located 12 km east of Olympia, Washington. Elevation is about 50 m. Climate is mild with an average growing season of 190 frost-free days and a mean July temperature of 16º C. Precipitation averages 1290 mm per year, falling mostly as rain from October through May; summers are periodically dry. Prior to installation of this study, the immediate area was in native forest occupied by Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco] and several hardwood species and shrubs. The trees were felled, merchantable stems removed, stumps pushed out of the ground, and the non-merchantable material was burned in piles. The topography is level and is occupied by two distinct soil types, both derived from glacial outwash. Most of the area contains a very deep, somewhat excessively drained, loamy sand (soil series Indianola, classified as mixed, mesic Dystric Xeropsamment). A minor portion of the area had gravel contents of 20 to 30% in the loamy sand surface soil, but does not drain as rapidly as the formerly mentioned soil type. Because of low rainfall during the growing season, neither soil would be considered suitable for commercial Populus plantations without irrigation.

The site was disked after harvest and burning, and a mix of N-P-K fertilizers was applied one month prior to planting to provide the equivalent of approximately 100 kg each of N, P, and K per hectare. In addition, 900 kg lime ha-1 was applied as a mixture of limestone and dolomite. Between the second and third growing season, 100 kg N ha-1 was applied as ammonium nitrate. Both the preplant and subsequent fertilizer applications (including lime) were spread on the soil surface but not incorporated. Preplanting and post-planting herbicide applications and hoeing were used as necessary to maintain the plots in a weed- free condition. Irrigation was provided by drip lines laid down 2 m apart; emitters (2.3 liters per hour) were spaced at 1-m intervals along each line. Amounts of water applied varied by year and weather conditions, ranging from 75 cm to 100 cm ha-1yr-1.

Experimental design and treatments

The study was established as a factorial design with five clonal treatments (four clones planted in monoclonal plots and one polyclonal plot with all clones in intimate mixture) and three square spacings, replicated in three blocks.

The four Populus clones were selected for use based on availability of stock, contrasting branching characteristics, and superior growth. Three of the clones were P. trichocarpa × P. deltoides hybrids developed in the University of Washington -Washington State University poplar breeding program (Heilman and Stettler 1985; Quinsey and others 1991); they have been used in many commerical plantations, and cuttings were provided by James River Corporation:

• Clone 11-11 is one of the first hybrids developed in the program, and it has been planted extensively throughout the Northwest; it grows rapidly and produces many sylleptic branches (i.e., branches that develop and elongate during the same growing season in which the bud is formed).
• Clone 47-174 grows very rapidly but produces very few sylleptic branches.
• Clone 49-177 grows very rapidly and produces many sylleptic branches.
• The fourth clone is a local selection of P. trichocarpa named Capitol Lake (CL) that had demonstrated excellent growth in small research plantings. It produces many sylleptic branches.

The three square spacings (0.5-m, 1.0-m, and 1.5-m) were selected to provide a range of stand density conditions; the two wider spacings have been commonly used in research plantings and in operational bioenergy plantations. The narrowest planting (0.5-m) provided a treatment in which competition developed more rapidly and to a higher degree.

Plot installation

Plot size varied by spacing; each treatment plot consisted of a 100-tree interior measurement plot (ten rows by ten columns) surrounded by three to eight buffer rows. Areas where debris piles had been burned were delineated and excluded from use in the study. Two blocks of plots were located on the major soil type; the third block was placed on the gravelly, less rapidly draining soil type.

The plots were established with unrooted woody cuttings which were >1 cm in diameter, 30 cm long, and had several healthy buds present. Cuttings were soaked in water overnight and then firmed into holes created with metal rods. The goal was to insert approximately 25 cm of the cutting length into the ground, but two healthy axillary buds were to remain above ground. Previous experience indicated establishment success (i.e., survival and early growth) was poor if cuttings did not have at least one healthy bud above ground (Radwan and others 1987). Requiring two buds above-ground ensured that a high percentage of cuttings sprouted but necessitated a later pruning to remove secondary or multiple stems. Planting was done during the last week of March 1990; stem pruning was done in autumn 1990.

Polyclonal plots were planted with 49-177 and CL alternating in even-numbered rows and 11-11 and 47-174 alternating in odd-numbered rows. Thus, the eight trees surrounding any individual subject tree represented a consistent composition of three clones, all of which differed from the subject tree.

Data collection and analyses

In both monoclonal and polyclonal plots, the 100-tree interior plot was used to measure tree dimensions and estimate standing biomass at the end of the third growing season. Tree diameter and height were recorded, and any unusual conditions (stem characteristics, stress or damage due to weather, insects, or diseases) were noted. Tree diameters were measured at 0.3 and 1.3 m above ground with metal diameter tapes and were recorded to the nearest 0.1 cm. Heights were measured with telescoping fiberglass poles and recorded to the nearest 5 cm.

Indices for lower-stem taper (0.3 m diameter/dbh × 100) and slenderness (ht/dbh × 100) were calculated from measurements of diameter and height. Coefficients of variation for diameter and height were calculated for each 100-tree measurement plot, and the three plot values were averaged to provide a mean coefficient of variation for each treatment.

To estimate standing biomass at age 3, five trees were selected from the buffer rows surrounding the interior measurement plot in each monoclonal plot. Trees in the outermost buffer row and in the buffer row immediately adjacent to the measurement plot were excluded. In total, 180 trees were selected to cover the range of tree sizes present in the study (i.e., 15 trees for each clone and spacing). Biomass-estimation trees were measured for height and stem diameter at 0.3 and 1.3 m, felled, and the number of live and dead branches counted. Live branches were classified as either sylleptic or proleptic. The trees were then divided into component parts (stem and live branches) and fresh weights of the components were determined. Subsamples of each were selected for determination of moisture content and dried to constant weight at 105º C.

Branching traits of the clones were examined using data from the two largest trees of the five trees sampled for biomass in each plot. Such trees were representative of trees in the stand's canopy. An index of branchiness was calculated as the percentage of total live woody weight associated with live branches. In addition, mean tree values were determined for numbers of live and dead branches, live branches only, and sylleptic branches formed in 1992.

Regression equations to predict dry biomass by component (stem or branches) were developed for each clone using logarithms of tree diameter at 0.3 m and height as forced independent variables. Coefficients of determination (R2) ranged from 0.957 (live branches of clone 49-177) to 0.997 (stems of clone 47-174). Spacing was also evaluated for inclusion as an additional variable, but it contributed significantly to only two equations (stem biomass of clone 49-177 and branch biomass of clone 47-174). Partial R2 for spacing in such instances was only 0.01. The equations were adjusted for log bias and then used to predict biomass components of individual trees in each permanent measurement plot. Stem and live branch weights were summed by component for each plot, and the sum was expanded to provide estimates of biomass per hectare.

Relative performance of each clone in polyclonal and monoclonal plantings was also evaluated by comparing the yield of the clone in polyclonal plots to its yield in pure culture, assuming an equal area occupied (or equal number of stems at time of planting). Thus, “comparative yields” of each of the four clones in polyclonal plantings were calculated for each spacing by dividing the yield of the clone in polyclonal plots by one- fourth of the yield it attained in monoclonal plots.

Three types of setups were used for analysis of variance (ANOVA). To answer most questions associated with individual trees, e.g., “Are there differences among clones, spacings, and deployment in mean tree size and stem form?”, the first-order sources of variation (with their degrees of freedom) were: blocks (2), clone (3), deployment (1), and spacing (2). Some branch data, however, were collected only in monoclonal plots; in these instances, the first-order sources of variation for the analyses were blocks (2), clone (3), and spacing (2). Both of these analyses used mean tree values by plot and clone as response variables. To answer the questions associated with plots, “Are there differences in yield or stand uniformity between monoclonal and polyclonal blocks?”, we used plot yield (i.e., the sum of the individual tree values converted to a per hectare basis) and coefficients of variation as response variables. For such analyses, there were five clonal deployment options or treatments (four monoclonal and one polyclonal). The ANOVA setup had clonal treatment (4), block (2), and spacing (2) as the first-order sources of variation. Survival data were analyzed using the logit transformation (Sabin and Stafford 1990).

Treatment effects were judged as significant when the probability of a greater F value was £ 0.05; however, the actual probability values are provided for readers to make judgements based on other values. Means were separated by Ryan-Einot-Gabriel-Welsh multiple F test procedures (SAS Institute Inc. 1988).

Results

The experimental design provided a sensitive assessment of treatment effects because tree growth was very rapid (2-4 m height increment per year) and environmental conditions within and among treatment plots and within blocks were extremely uniform. The main effects of clone and spacing were significant for all traits, deployment was significant for tree size and lower-stem taper, and interactions between two or three main effects were significant for most individual tree characteristics but not for stand characteristics such as yield and size variation (Table 1). Effects of block were significant for all tree and stand traits related to growth or productivity (but not to stem form, branching habit, or tree size variation); as expected, the third block on the gravelly, more poorly drained soil had smaller trees and lower yields than the other two blocks which were similar.

Survival

Initial establishment success (i.e., root and shoot development of cuttings) was excellent. All planting positions were occupied with vigorous trees by early summer and all had a surviving tree at the end of the first growing season. Competition-related mortality began to occur in the 0.5-m spacing during the second year, most of which occurred in the CL clone planted in the polyclonal plots. In other clones, at least one tree died at this spacing in both monoclonal and polyclonal plots. Clones 47-174 and 49-177 suffered some mortality at all spacings due to an unidentified shoot blight. By the end of the third- growing season, survival averaged 92% and differed significantly by clone, spacing, and their interaction with deployment. The vast majority of the mortality, however, occurred in the CL clone planted at 0.5-m spacing in polyclonal blocks; only 43% of trees remained (Table 2). The other combinations of clone and deployment in the 0.5-m spacing had much higher survival (81 to 97%). At the 1.0-m and 1.5-m spacings, CL, 11-11, and 47-174 had 96 to 100% survival. Clone 49-177 had slightly lower survival (88 to 93%) at these spacings, and all of the mortality was associated with an unidentified shoot blight.

Tree size

Height and diameter varied significantly with spacing, clone, manner of deployment, and interactions thereof (Table 3). Averaged over all clones in monoclonal plantings, height increased from 6.8 to 11.0 m, and diameter increased from 3.3 to 7.5 cm as spacing widened from 0.5 to 1.5 m. Mean heights and diameters for each clone averaged over all spacings in monoclonal plots were 9.2 m and 5.5 cm for 11-11; 9.3 m and 5.4 cm for 47- 174; 9.3 m and 5.8 cm for 49-177; and 8.8 m and 5.2 cm for CL. Size differences among the clones in the monoclonal plots increased as spacing increased; the range in height increased from 0.3 m at 0.5-m spacing to 0.7 m at 1.5-m spacing, whereas the range of diameters increased from 0.3 cm to 1.0 cm over the same spacings. Although actual rankings of clones in terms of mean height and diameter varied only minimally with spacing in the monoclonal plantings, absolute and relative differences among clones were substantial in polyclonal plantings, and a shift in ranking by height occurred between clones 11-11 and 47-174. In monoclonal plantings, clone 47-174 was slightly taller than clone 11-11, whereas the latter clone was slightly larger in diameter. In polyclonal plantings, however, clone 11-11 was larger than clone 47-174 in both height and diameter, and differences between the clones became significant as spacing widened to 1.5 m (Table 3). Clone CL was the smallest clone in all monoclonal plantings, and its size relative to the other clones was diminished dramatically in polyclonal plantings.

Tree form

Lower stem taper and slenderness differed among clones and spacings, and these traits differed in some clones and some spacings between monoclonal and polyclonal plots (Table 4). At age 3, clone 11-11 had the least taper; its taper did not change significantly with spacing and did not differ in monoclonal vs. polyclonal plantings. Clone 47-174 had the greatest lower stem taper in monoclonal plots and differed significantly from other clones at the two widest spacings, but its taper did not differ between polyclonal and monoclonal plantings. The taper of clones 49-177 and CL tended to be intermediate between the other two clones. Taper of clone CL did not differ significantly with spacing in monoclonal plots, but it decreased with increased spacing in polyclonal plots. Moreover, the taper of clone CL was significantly greater in polyclonal than in monoclonal plantings at the two closest spacings. Taper of clone 49-177 was unaffected by either spacing or deployment.

Slenderness differed minimally among clones in monoclonal plantings, but clonal differences were substantial in polyclonal plantings where slenderness of clone CL was significantly greater and that of clone 49-177 was substantially less than the other two clones. Slenderness decreased with increased spacing as expected because diameter growth is enhanced to a greater degree than height growth by increased growing space. Deployment had only minor effects on slenderness of clones 11-11 and 47-174; effects on the other two clones were more substantial and occurred in opposite directions for each. Slenderness of clone 49-177 was lower in polyclonal plots than monoclonal plots, whereas that of clone CL was significantly greater in polyclonal plots.

Branchiness as defined by the percentage of total live woody weight due to live branches differed significantly by clone, spacing, and their interaction. On average, clone CL had the highest branch index which was significantly greater than that of clone 11-11; branchiness of other clones was intermediate and did not differ from either clone CL or clone 11-11. Branch biomass index increased with increased spacing, being significantly greater at 1.5- m spacing (11%) than at the two closest spacings (5 and 7%). Ranking of the clones, however, changed markedly and clonal differences were greater for numbers of branches (Table 4). Total branches (live and dead) increased significantly as spacing widened from 0.5 m (80 branches) to 1.0 m (112 branches) and further increased, though to a lesser, non-significant extent, in the 1.5-m spacing (119 branches). Clone 47-174 produced significantly fewer branches (67) than any other clone, and CL had the most branches (119). Differences among clones were even greater for live branches and 1992 sylleptic branches. Capitol Lake had the greatest number of live branches (59) and clone 47-174 the least (26 branches), with the other two hybrid clones (11-11 and 49-177) having similar amounts (42 branches each). Live branch numbers increased with increased spacing, with the number in each spacing differing significantly from the other two spacings. Number of sylleptic branches was reduced at close spacing and differed more among clones than other branch traits. Clones 11-11 and CL produced similar numbers of sylleptic branches (24 and 28, respectively); 47-174 had none in the closest spacing and averaged only one branch per tree in the two wider spacings; and clone 49-177 produced intermediate numbers ranging from 8 in the closest spacing to 24 in the widest spacing.

Stand uniformity

Clonal plantations established on intensively prepared sites are relatively uniform compared with most naturally regenerated forests. Nevertheless, stand differentiation does occur; at the end of the third year in this study, there were significant differences in height and diameter differentiation (as measured by coefficient of variation) across spacings and among clonal treatments. On average, the 0.5-m spacing exhibited the greatest within-plot variability in heights and diameters and the 1.5-m spacing showed the least (Table 5). The hybrid clones were somewhat more variable than clone CL in the monoclonal plantings; among the hybrids, clone 49-177 tended to be the most variable, especially at the two wider spacings. Coefficients of variation for various traits in polyclonal plots were higher than the mean for the monoclonal plots at the same spacing, and with the exception of the height variation of clone 49-177 in 1.5-m spacing, they were significantly higher than for any clone in corresponding monoclonal plots (Table 5). Although coefficients of variation for all clonal treatments were greatest in the 0.5-m spacing, relative differences between polyclonal and monoclonal plots in within-plot variation were smallest for diameter and height at that spacing and increased with increased spacing. For example, coefficients of variation for diameters in the 1.0-m and 1.5-m polyclonal plots were more than 60% higher than the mean of coefficients of variation for the corresponding monoclonal plots, whereas they were only 22% higher at the 0.5-m spacing. Similarly, polyclonal plantings had coefficients of variation for diameter that were 33 to 41% higher than those for the most variable clone (49-177) in monoclonal blocks at 1.0- and 1.5-m spacing and only 13% higher than the most variable monoclonal plantings (11-11) at 0.5-m spacing. In general, trends for variation in height are similar (though less striking) to those for variation in diameter.

Stand yield at age 3

Above-ground yields differed significantly among clones and spacings (Table 6). Three- year total live woody yields ranged from a low of 35.2 Mg ha-1 for clone CL at 1.5-m spacing to a high of 54.9 Mg ha-1 for clone 49-177 at 0.5-m spacing. Averaged over all spacings, total live woody yields in Mg ha-1 at age 3 in monoclonal plantings were 48.7 for 49-177, 45.9 for 11-11, 45.3 for 47-174, and 37.3 for CL. Total woody yields of polyclonal plots (43.1 Mg ha-1) were significantly higher than yields from monoclonal plots of CL. Total woody yields of all clonal treatments decreased as spacing increased, with yield at 1.5-m spacing being significantly lower than yields at the two closer spacings. Stem yield patterns were similar to patterns for total live woody yields as they constituted more than 90% of total yield (Table 6). Branch yield, however, increased with increased spacing and was significantly greater at 1.5-m spacing than at the other two spacings. Clonal rankings in branch yield also differed from rankings in stem and total woody yield; overall, clone 49-177 had significantly higher branch yield than the other three clones in monoclonal plots. At the widest spacing, however, branch yields of clones 47-174 and CL were slightly higher than those of clone 49-177, a striking reversal of the clonal rankings of branch yield at the 0.5-m spacing.

Both stem and total woody yield of polyclonal plots tended to be slightly lower than those of the average of monoclonal plots but not significantly so (Fig. 1, Table 6). Averaged over all clones and spacings, monoclonal plots yielded 44.3 Mg ha-1 of live woody biomass whereas polyclonal plots yielded 43.1 Mg ha-1. There were substantial differences, however, between monoclonal and polyclonal deployment in the contribution of each clone to total yield (Fig. 1). Moreover, the magnitude of such differences varied by spacing. The clones made rather similar contributions (22 to 27%) to total yield in monoclonal plantings at the widest (1.5-m) spacing. For polyclonal plantings at that spacing, clone 49-177 and clone 11-11 provided 35% and 31% of total yield, respectively, whereas clones 47-174 and CL provided only 20% and 14%. The disparity among clones in polyclonal plots was much greater as spacing decreased. At 0.5-m spacing, clone CL provided only 1%, whereas clone 49-177 provided 46% of the total yield.

Figure 1. Total biomass per hectare at age 3 produced by four Populus clones deployed in monoclonal and polyclonal planting at three spacings; assuming equal numbers of trees were initially planted for each clones in both methods of deployment.

Comparative yield

Comparative yield — that is, clonal yield values in polyclonal plots expressed as a fraction of those obtained in monoclonal plots, assuming equal area is occupied by the clone — provides additional clarification of clonal interactions in mixed plantings. Whether one looks at total live woody biomass or stem (not shown), it is apparent that each clone manifested a different response to mixture and spacing treatments (Fig. 2). Clone 11-11 yielded slightly more in polyclonal than in monoclonal plantings with the greatest difference (+18%) occurring at the widest spacing. Clone 49-177, on the other hand, grew markedly better (+58%) in polyclonal plantings than in monoclonal plantings at 0.5-m spacing. But as spacing widened, the superiority of its growth in polyclonal plots relative to monoclonal diminished and was only 26% better at 1.5-m spacing. In contrast to the above two clones which had enhanced growth in polyclonal plots, yields of clones 47-174 and CL were lower in polyclonal than in monoclonal plots; the extent of the decreases and trends with spacing differed markedly, however, between the latter two clones. Clone 47- 174 was less affected and the relationship with spacing was nonlinear. Mild detrimental effects of the mixture on relative yield at 0.5-m spacing (12% lower than in monoclonal plots) became nearly negligible (about 3% lower) at 1.0-m spacing and then increased at 1.5-m spacing (-22 to -25%). Yield of clone CL was severely depressed in polyclonal plantings at 0.5-m spacing, providing less than one-tenth the yield in monoclonal plots. Detrimental effects, however, became less severe with increased spacing (about 40% lower at 1.5-m spacing).

Fig 2. Comparative yields (polyclonal + monoclonal) of four Populus clones with polyclonal plantings as related to spacing. Values exceeding 1.0 indicate highter productivity of the clone in polyclonal plantings; values lower than 1.0 indicate reduced productivity.

Discussion and Conclusions

By age 3, wider spacings produced trees that were taller and had larger diameters than were produced in closer spacings. Due to fewer number of trees per unit area, however, narrow spacings maximize biomass production during the initial years after planting. The wider the spacing, the more uniform were tree diameters and heights. There were only minor changes in clonal rankings for most tree and stand characteristics across spacings in monoclonal plots, but the advantage in yield associated with being the top ranked clone decreased substantially with increasing spacing. Thus, if evaluations were made at a narrower spacing than would be used operationally, the yield advantage could be substantially overestimated. Differences in relative performance of clones across spacings can be attributed to: (1) clonal branching habits and the effects of spacing on the expression thereof and (2) other clonal differences in physiological responses to self- shading and other aspects of intra-clonal competition.

Clonal differences and spacing by clone interactions were enhanced in polyclonal plantings. Moreover, differences in relative performance of clones in polyclonal vs. monoclonal plantings increased with time. Clones such as CL that grew more slowly than others in monoclonal plots would be expected to grow even more slowly in polyclonal plots once competition reached levels detrimental to growth. In polyclonal plantings, the good clones tend to get better and the poor clones get poorer. But even clones that are superior in monoclonal plots (such as 47-174) may be placed at a competitive disadvantage when faced with neighbors of differing physiological or morphological attributes. Such relative differences among clones in polyclonal vs. monoclonal plantings also generally increased with stand density (i.e., they were greater at 0.5-m than at wider spacings). Interactions between density and deployment strategy were very striking for clones CL and 49-177. On the other hand, the performance of clone 11-11 was little affected by deployment strategy or its interaction with spacing. And clone 47-174 again provides an exception to the generality for some traits; for example, its yields in polyclonal and monoclonal plantings were more similar at 0.5-m than at 1.5-m spacings. The latter reversal probably was related to the branching habit of 47-174; reduced sylleptic branching was not as great a disadvantage at close spacing where syllepticity of all clones was minimal. At the wider spacing, however, where syllepticity was fully expressed, clone 47- 174 was at a much greater competitive disadvantage relative to other clones such as 11-11 and 49-177 which grew rapidly and produced many sylleptic branches.

Some people argue on theoretical grounds that yield in polyclonal plantings may be higher and that deployment in more diverse plantings may reduce risks by protecting populations from catastrophic losses. Our study failed to show any yield advantage of polyclonal plantings. Monoclonal yields of some individual clones exceeded polyclonal yields, but not significantly so; moreover, on average, polyclonal yields were slightly (though not significantly) lower than the 4-clone average of monoclonal yields. Hazards that may hinder tree survival and growth are many as are the mechanisms through which they enter, affect, and spread through a plantation. Such differences in damaging agents are obviously important considerations in deployment strategies. Other things being equal, however, the theoretical risk-spreading advantages may be less than one might otherwise assume if relative yield of individual clones changes markedly in polyclonal plantings. In 0.5-m spacings, for example, 46% of inventory was tied up in just one clone (49-177) and another 29% in a second clone (11-11); the remaining two clones accounted for only 25% of inventory. When the same four clones were deployed in monoclonal plantings, however, their relative contribution to overall production and inventory was much more similar, ranging from 20% for clone CL to 28% for clone 49-177. In our plantings, risks were spread over a less balanced inventory — in effect, relying on a less diverse population — when the same four clones were deployed in polyclonal plantings than in monoclonal plantings. Although this effect might be reduced with inclusion of additional or different clones, the same principle may apply. It therefore seems important to understand and consider the effect of deployment strategies on the distribution and balance of inventory among clones.