Short and long-term effects of treeshelters on the root and stem growth of ornamental trees

D.W. Burger1, G.W. Forister1 and R. Gross2

1 Department of Environmental Horticulture, University of California, Davis, CA 95616

2 DendroTech, Calistoga, CA 94515

Additional index words. biomass, partitioning, dry matter, hydraulic soil excavation

Abstract. Short-term (aerated solution culture and container nursery) and long-term (landscape) experiments were conducted to study the effect of treeshelters on the root and stem growth of several ornamentals (Sequoia sempervirens (D. Don) Endl., Quercus lobata Née, Quercus agrifolia Née, Lagerstroemia indica 'Watermelon Red', Ginkgo biloba L., Platanus racemosa Nutt., Fraxinus latifolia Benth., Dendranthema grandiflora 'Bright Golden Anne' and Populus euamericana cv. 'Giacometti'). In general, plants grown in treeshelters were taller and some had reduced caliber growth. When grown in shelters, top dry weight was reduced for Fraxinus latifolia, Platanus racemosa, Quercus agrifolia, Quercus lobata, Dendranthema grandiflora and Populus euamericana compared to unsheltered plants. Root dry weight, root:shoot ratio, total root length and total root area were reduced for all plants except Quercus agrifolia when grown in shelters. The results are explained on the basis of the microenvironment in/around treeshelters, photosynthetic partitioning and immobilization of plants growing in shelters. Management challenges and potential usefulness of treeshelters in landscape transplanting is also discussed.

Treeshelters have been used to help improve transplantation success of trees in the landscape. Work around the world over the past decade has shown that treeshelters increase survival (Costello et. al., 1996), enhance growth rates and tree height with relatively little effect on stem caliber and overall biomass (Burger, et. al., 1992; Frearson and Weiss, 1987; Potter, 1988). Relatively little attention has been paid to the development of roots or dry matter partitioning of trees growing in shelters. Rendle (1985) found that Quercus robur trees growing in shelters grew to set heights in shorter periods of time, but did not differ in total dry matter accumulation; however, she did find that dry weight between stems, branches and roots differed depending on whether treeshelters were used or not. More recently, vihra et al. (1996) found that treeshelters reduced the fresh and dry weight of redwood (Sequoia sempervirens) trees growing in the landscape for 4 years.

Dry matter partitioning can be an important factor in the early development of trees in the landscape. If root development is inhibited in some way before or during transplantation the newly planted tree has a lower probability of becoming well established (Harris, 1992). The objective of this two-year study was to determine the effect of treeshelters on the root development and dry matter partitioning of seven tree species common to the western landscape.

Materials and Methods

Short-term experiments. A randomized, complete block design experiment (6 blocks, 1-2 replicates per block) was initiated in May, 1993, in a container nursery whereby each block contained two replicates of each treatment, trees with shelters (122-cm tall, tan Tubex® shelter) and those without (control). Seed-propagated liners were used as planting stock including: Sequoia sempervirens (D. Don) Endl., Quercus lobata Née, Quercus agrifolia Née, Lagerstroemia indica 'Watermelon Red', Ginkgo biloba L., Platanus racemosa Nutt. and Fraxinus latifolia Benth. The trees were grown in 1.8-liter (#1) containers filled with UC Mix (1:1:1, by volume, sand:redwood sawdust:peat moss). Thirty g of Nutricote Total (18-6-8 with micronutrients, Type: 180, PlantCo, Inc., Brampton, Ontario, Canada) were added to the soil surface at planting time. Treeshelters were placed over the appropriate plants after the containers were moved to the nursery bed. The nursery containers were placed on 40-cm centers. An automated irrigation system (spray stakes) was used to add approximately 500 ml of water each day. Half the trees in each block were harvested after three months and the other half after six months. Measurements of height, caliber and fresh/dry biomass accumulation were taken at each harvest.

Long-term experiments. A randomized, complete block design landscape experiment (10 blocks) was initiated in April, 1993, whereby each block contained two adjacent trees of each of seven tree species. One tree in each pair was grown in a treeshelter. The same seven tree species were used as in the nursery experiments, but were planted on 1.8 m centers in rows 1.8 m apart. The trees were irrigated with an automatic drip system that provided ample water for the duration of the experiment. Trees were fertilized once, four weeks after planting, with Agriform 20-10-5 plus minors fertilizer tablets (Grace/Sierra, Milpitas, CA). The fertilizer tablets were pressed 8-12 cm into saturated soil immediately after an irrigation. After 12 months half (n=10 for each treatment) the trees' height and stem diameter (caliber) were measured and fresh and dry weights were determined for tops (leaves, branches, stems) and roots. Root systems were extracted from the soil with a backhoe fitted with a 60-cm wide bucket. The soil around the root systems was removed with pressurized water so root fresh and dry weights could be determined. For fresh weight determinations, recently rinsed root systems were shaken dry and weighed (fresh weight). The root system was then cut into pieces, placed in a paper bag, dried for at least seven days at 70C and weighed (dry weight). The remaining trees were harvested after 24 months. Trees were harvested by cutting the trees at ground level. Leaves and stems were cut into pieces, weighed (fresh weight), placed in paper bags and dried for at least one week at 70C to obtain dry weights. Collected biomass data included top (leaves, branches, stems) fresh and dry weight. Root systems were exposed using Hydraulic Soil Excavation (Gross, 1993; 1995). A pneumatic conveyor industrial loader vacuum truck (Vactor model 1645, Federal Signal Co., Streator, IL) removed the resulting slurry. The water was supplied by a nearby hydrant with a flowing pressure of 64 PSI. The water was delivered from the hydrant with a 2- in. diameter supply line. The hose size was reduced to 1- in. diameter handline with 1- in. fittings at the excavation site. A modified "EZ Pup" (Durham Fire and Rescue Specialty Tools, North Tonawanda, NY) was utilized at the end of the handline to reduce operator fatigue. A KK "Thunder fog" nozzle (Task Force Tips, Valpariso, IN) was attached to the "EZ Pup". The water stream delivered under pressure reduced the soil in the root area to a slurry and the Vactor vacuum hose removed it. Using this technique, undamaged roots as small as 1 mm in diameter were exposed. On deeper, more extensive root systems, hydroexcavation continued until roots as small as 1 cm were exposed. They were then cut at the soil interface before removing the excavated portion from the ground. Once extracted, the root systems were air-dried, painted white, suspended above a back-lit background and photographed from above using high-contrast black-and-white film. This set up provided excellent contrast between the shadow-darkened root system and the bright background. Once photographed the root systems were cut into pieces and dried for at least seven days at 70C (for dry weight). The black-and-white photographs were used to estimate total root length and total root area. Each root system was photographed with a rod of known length (10, 50 or 100 cm) in the picture. The images of known length and a piece of string arranged into a circular shape (ends not touching) were used to calibrate the video imaging system (AgVision Root and Leaf Analysis, Decagon Devices, Inc.) that made the root length and root area estimations. The string was at least 150 mm in length and produced an image at least 200 pixels long on the computer screen to accurately calibrate the digitized image. By comparing the known length of the string and the known length of the rod in each photograph, the actual dimensions of the root system could be determined by the AgVision computer software.

Solution culture experiments. Dendranthema grandiflora 'Bright Golden Anne' and Populus euamericana cv. 'Giacometti' were grown in aerated solution cultures in a greenhouse (22 to 28C). Tip cuttings of Dendranthema and Populus were rooted in aero-hydroponic units (Soffer and Burger, 1988) containing de-ionized water. Once rooted, the cuttings were put singly into containers holding approximately 2 liters of Hoagland's Solution 2 plus micronutrients (Hoagland and Arnon, 1950) that were continuously aerated (1 litermin-1). Each cutting was held in place in the lid of the container with rubber stoppers that had holes bored through them. This arrangement allowed the rooted cuttings to be held with only their root systems immersed in the aerated nutrient solution. Each container was equipped with an aeration tube, a drip tube capable of distributing fresh Hoagland's Solution 2 and an overflow tube (just under the lid). Treeshelters were hot-glued to the top of half the containers; care was taken to insure there was a seal between the treeshelter and the container top so no air movement could occur there. There were six containers per treatment arranged in a completely randomized design on the greenhouse bench. Every day, fresh nutrient solution was added to each container using an automated irrigation system. When the container became filled during each "irrigation", excess nutrient solution escaped through the overflow tube. Plants were measured and harvested four weeks (Dendranthema) and six weeks (Populus) after placing the rooted cuttings in the nutrient solution containers. All data for all experiments were analyzed using the General Linear Model (GLM) Procedure of the SAS statistical system (SAS Institute, 1988).

Results

Height increase. All species growing in shelters except G. biloba showed height increases after only three months (May to August) of growth (Table 1). At that time height for sheltered trees (excluding G. biloba) was enhanced 63-244%. This same result occurred in those tree species measured at six months (Table 1). After one year, L. indica, Q. agrifolia and Q. lobata trees growing in shelters were still taller than their unsheltered counterparts, whereas F. latifolia, G. biloba, P. racemosa and S. sempervirens trees growing in shelters were no taller than their unsheltered counterparts. After two years only L. indica and S. sempervirens trees growing in shelters continued to be taller than the controls (Table 1). The unsheltered trees of the other five species all had similar heights by then.

Caliber development. Caliber development was only slightly affected by treeshelters after three months in F. latifolia, G. biloba and S. sempervirens (Table 2). At six months none of the tree species showed differences in caliber and long-term (1 or 2 years) effects of treeshelters on caliber were only observed in F. latifolia (after two years), P. racemosa, Q. agrifolia (2 years) and S. sempervirens (after one and two years) (Table 2). Caliber differences between control (unsheltered) and sheltered trees were not observed at any time for L. indica or Q. lobata.

Top dry weight. Top dry weight (TDW) was not consistently affected by treeshelters during the time of the experiment among most of the species. After 6 months L. indica, Q. agrifolia, Q. lobata and S. sempervirens trees growing in shelters had greater top dry weights than control trees (Table 3). However, after 1 year no tree species growing in shelters had higher top dry weight than control trees and P. racemosa and S. sempervirens unsheltered trees had higher top dry weights than sheltered counterparts (Table 3). After 2 years F. latifolia, P. racemosa, Q. agrifolia and Q. lobata trees growing in treeshelters had lower dry weights than control trees.

Root dry weight and dry matter partitioning. Except for Q. agrifolia, all tree species eventually had reduced root dry weights (RDW) when grown in treeshelters (Table 4). The reduction in RDW occurred most rapidly (3 months) for F. latifolia, G. biloba and P. racemosa and by six months for L. indica, Q. lobata and S. sempervirens. After two years, RDW of sheltered trees ranged from 9 to 71% that of unsheltered trees. The root:shoot ratio was reduced in sheltered trees of several species including G. biloba, L. indica, Q. lobata and S. sempervirens (Table 4). Most dramatic was the response observed in L. indica and S. sempervirens where the root:shoot ratios for the sheltered trees were 11% and 8% that of the controls.

Total root length and total root area. Total root length and total root area were both lower in sheltered versus control trees for all tree species except for Q. agrifolia (Table 5). Total root length was reduced from 17-39% and total root area was reduced from 38-89% when trees were grown in shelters.

Dry biomass response of Dendranthema and Populus in aerated solutions. Leaf, stem and root dry weights and root-shoot ratios of Dendranthema and Populus were lower in plants growing in shelters versus those unsheltered (Table 6). The percent reductions were similar in these two species growing in aerated nutrient solutions compared to tree species growing in containers and in the landscape. In these experiments none of the root systems were lost when the plants were harvested compared to the trees hydroexcavated from landscape sites. Height and fresh weight responses were also similar to those of plants growing in soil (data not shown).

Discussion

The most striking response observed in this study was that of the reduced root biomass from trees growing in treeshelters. Six of the seven tree species growing in soil and both species growing in aerated solution cultures showed reduced root biomass, total root length and total root area. Some roots from very large root systems associated with control (unsheltered) trees in the landscapes sites were left unexcavated due to their depth and lateral expanse. This means that growth estimates of those root systems are conservative. Still, significant differences were found between sheltered and unsheltered treatments for most of the tree species studied. The reduction in root biomass was the result of a reduced overall photosynthate pool (reduced biomass production) and/or a change in photosynthate partitioning between treatments causing a reduction in the root:shoot ratio. Root biomass and root:shoot ratios of trees have been shown to be adversely affected by water and nutrient regimes (Fabiao et al., 1995), planting density (Puri et al., 1994) and light (Dias-Filho, 1995), but this is the first report of a long-term reduction in root biomass and root:shoot ratios in a variety of species in response to treeshelters. The reduced root biomass found in most of the tree species in this study is most likely due to a combination of environmental factors associated with treeshelters. The light intensity inside the treeshelter is approximately half that outside the treeshelter (Burger et al., 1992 ). This alone would reduce overall available photosynthate.

Top dry weight changes were due either to changes in height or changes in caliber. In no case were height increases associated with caliber increases and in one case (S. sempervirens) increases in height were accompanied with decreases in stem caliber. This is a common occurrence in woody perennials as available photosynthate is partitioned into competing above-ground sinks.

The light environment surrounding a plant shoot can affect its root size and root/shoot biomass ratio and has been widely studied. Individual plant responses are likely under genetic control and "strategies" have been suggested by which a plant invests only enough carbon in roots to support the plant as is develops. Kasperbauer and Hunt (1992) found in sweetclover (Melilotus alba Desr.) that while photosynthesis regulates carbon fixation and biomass production, photoperiod and spectral distribution regulate biomass distribution (partitioning) and new photosynthate. Wilson (1988) has also provided a summary of the effects of light and carbon dioxide on the root:shoot ratios of plants. Changes in light, in particular in light quality or spectral distribution, could explain why root:shoot ratios differ between sheltered and unsheltered tree species studied here. Tubex® treeshelters used in these experiments were tan in color and have been shown to reduce light intensity. While spectral shift analyses have not been conducted for these treeshelters, it is quite probable that the absorption, transmission and/or reflection of light inside the shelter is substantially different from incident light outside. Analyses such as these will be necessary to support this contention and help provide an explanation for the dramatic root biomass reductions observed in most trees growing inside treeshelters.

Not all trees responded similarly to treeshelters. For some, height was influenced early and diminished over time (e.g., F. latifolia, P. racemosa and Q. lobata); others were affected for longer periods of time (in this study, two years) (e.g., L. indica and S. sempervirens); and others, were affected very little (e.g., G. biloba). For most tree species in this study, as the tree grew and eventually emerged from the 122-cm treeshelter, the influence the shelter had on growth of the tree diminished. This is to be expected especially when trees had more than half their above-ground biomass outside the treeshelter (e.g., F. latifolia, P. racemosa and Q. lobata). For other tree species this was not the case. L. indica and S. sempervirens trees maintained their differences in height even though nearly 50% of the tree was out of the treeshelter. There may be other physiological influences of the treeshelter on lower trunk (that part inside the treeshelter) that have not yet been identified or the early growth of these tree species inside the shelter may predispose them to longer term growth enhancements.

The responses observed in these tree species are similar to those seen by our earlier work and those of others (Burger et al., 1992, 1996; Frearson and Weiss, 1987; Potter, 1988, 1991; vihra et. al., 1996). That is, trees grown in shelters are taller, tend to have somewhat smaller stem caliber and have reduced fresh and dry weight of above and below-ground biomass. Environmental conditions (reduced light intensity, increased temperature, relative humidity and CO2 concentration) inside treeshelters can explain most of these growth responses. Treeshelters should not be left around trees once the tree has grown taller than the shelter. This especially important in windy areas since trunk deformities result (Burger et al., 1996). The benefits of treeshelters include enhanced stem height, protection from browsing animals and management practices (herbicide applications) and convenient release of predators inside to control of insect pests. The reduction in root mass may not always be a liability to the early development of newly transplanted trees in the landscape depending on the species.

Literature Cited

Burger, D.W., G.W. Forister and P.A. Kiehl. 1996. Height, caliber growth and biomass response of ten shade tree species to treeshelters. J. Arboriculture (in press)

Burger, D.W., P. vihra and R. Harris. 1992. Treeshelter use in producing container-grown trees. HortScience 27(1):30-32.

Costello, L.R., A. Peters and G.A. Giusti. 1996. An evaluation of treeshelter effects on plant survival and growth in a Mediterranean climate. J. Arb. (in press)

Dias-Filho, M.B. 1995. Physiological responses of Vismia guianensis to contrasting light environments. Revista Brasileira de Fisiologia Vegetal 7(1):35-40.

Fabiao, A. 1995. Development of root biomass in a Eucalyptus globulus plantation under different water and nutrient regimes. Plant and Soil 168-169:215-223.

Frearson, K. and N.D. Weiss. 1987. Improved growth rates within treeshelters. Quart. J. For. 81(3):184.187.

Gross, R. 1993. Hydraulic soil excavation: Getting down to the roots. Arbor Age 13:10, 12-13.

Gross, R. 1995. Hydraulic soil excavation. In: Trees and Building Sites, Eds. G.W. Watson and D. Neely. International Society of Arboriculture, Savoy, IL. pp. 177-184.

Harris, R.W. 1992. Arboriculture: Integrated management of landscape trees, shrubs and vines. Prentice Hall, Englewood Cliffs, NJ. 674 pp.

Hoagland, D.R. and D.I. Arnon. 1950. The water-culture method for growing plants without soil. Calif. Agr. Expt. Sta. Circ. 347.

Kasperbauer, M.J. and P.G. Hunt. 1992. Root size and shoot/root ratio as influenced by light environment of the shoot. J. Plant Nutrition 15(6-7):685-697.

Mayhead, G.J. and T.A.R. Jenkins. 1992. Growth of young Sitka spruce (Picea sitchensis (Bong.) Carr.) and the effect of simulated browsing, staking and treeshelters. Forestry 65(4):453-462.

Neel, P.L. and R.W. Harris. 1971. Motion-induced inhibition of elongation and induction of dormancy in Liquidambar. Science 173:58-59.

Potter, M.J. 1988. Treeshelters improve survival and increase early growth rates. J. Forestry 86(6):39-41.

Potter, M.J. 1991. Treeshelters. Forestry Commission Handbook 7. HMSO Publications Centre. P.O. Box 276, London, SW8 5DT.

Puri, S., V. Singh, B. Bhushan and S. Singh. 1994. Biomass production and distribution of roots in three stands of Populus deltoides. Forest Ecology and Management 65(2-3):135-147.

Rendle, E.L. 1985. The influence of treeshelters on microclimate and the growth of oak. Proc. 6th Natl. Hardwoods Prog., Oxford Forestry Institute.

SAS Institute. 1988. SAS/STAT user's guide, release 6.03 ed. SAS Institute, Cary, N.C.

Soffer, H. and D.W. Burger. 1988. Plant propagation using an aero-hydroponics system. HortScience 24(1):154.

vihra, P., D.W. Burger, and R.W. Harris. 1996. Treeshelter effect on root development of redwwod trees. J. of Arb. (in press)

Wilson, J.B. 1988. A review of evidence on the control of shoot-root ratio, in relation to models. Ann. Bot. 61:433-449.


Table 1. Mean height increases of seven tree species over time growing with (+) and without (-) treeshelters. (n=8 for 3- and 6-month data, n=20 for 1-year data, n=10 for 2-year data)

Species

Mean Height Increase (cm)
3 months
6 months
1 year
2 years
-
+
-
+
-
+
-
+
F. latifolia 27 b 44 a --- --- 162 a 197 a 238 a 233 a
G. biloba 0 a 0.3 a 1.5 a 4.4 a 13 a 18 a 34 a 51 a
L. indica 27 b 76 a 30 b 76 a 30 b 137 a 98 b 160 a
P. racemosa 38 b 64 a --- --- 134 a 173 a 330 a 308 a
Q. agrifolia 16 b 51 a 27 b 78 a 57 b 144 a 172 a 191 a
Q. lobata 9 b 31 a 19 b 53 a 71 b 200 a 233 a 261 a
S. sempervirens 11 b 25 a 12 b 38 a 85 a 68 a 123 b 213 a

Values followed by the same letter are not significantly different from one another at p=0.05 using Scheffe's multiple separation procedure.


Table 2. Mean caliber increases of seven tree species over time growing with (+) and without (-) treeshelters. (n=8 for 3- and 6-month data, n=20 for 1-year data, n=10 for 2-year data)

Species

Mean Caliber increase (mm)
3 months
6 months
1 year
2 years
- + - + - + - +
F. latifolia 9 a 6 b --- --- 32 a 27 a 83 a 61 b
G. biloba 2.0 a 1.3 b 2.6 a 2.2 a 3.5 a 3.2 a 9.1 a 6.4 a
L. indica 4.9 a 4.3 a 5.4 a 5.9 a 12 a 10 a 32 a 23 a
P. racemosa 7 a 9.0 a --- --- 57 a 36 b 138 a 82 b
Q. agrifolia 3.4 a 3.1 a 4.1 a 5.0 a 13 a 12 a 31 a 22 b
Q. lobata 1.8 a 2.4 a 2.1 a 2.7 a 16 a 12 a 44 a 32 a
S. sempervirens 3.2 a 1.6 b 4.5 a 5.2 a 24 a 12 b 52 a 36 b

Values followed by the same letter are not significantly different from one another at p=0.05 using Scheffe's multiple separation procedure.


Table 3. Top dry weight of seven tree species over time growing with (+) and without (-) treeshelters. Dry weight data for 3 months was lost due to a fire in the drying oven. (n=8 for 3- and 6-month data, n=20 for 1-year data, n=10 for 2-year data)

Species

Top Dry Weight (g)
6 months 1 year 2 years
- + - + - +
F. latifolia --- --- 542 a 463 a 3286 a 2265 b
G. biloba 3.7 a 3.3 a 11.4 a 6.8 a 23.2 a 17.0 a
L. indica 11.7 b 24.3 a 120 a 177 a 625 a 638 a
P. racemosa --- --- 1518 a 720 b 9357 a 3280 b
Q. agrifolia 22.6 b 48.2 a 113 a 143 a 677 a 533 b
Q. lobata 7.0 b 12.0 a 162 a 96 a 1346 a 925 b
S. sempervirens 16.5 b 31.3 a 386 a 102 b 1125 a 1288 a

Values followed by the same letter are not significantly different from one another at p=0.05 using Scheffe's multiple separation procedure.


Table 4. Root dry weight over time and final root-shoot ratios of seven tree species growing with (+) and without (-) treeshelters. Some roots <1 cm in diameter were remained in the soil after root system excavation. (n=8 for 3- and 6-month data, n=20 for 1-year data, n=10 for 2-year data)

Species

Root Dry Weight (g) Root:Shoot

(dry weight basis)

3 months 1 year 2 years
- + - + - + - +
F. latifolia 46.3 a 17.8 b 259 a 147 b 1473 a 1051 b 0.45 a 0.46 a
G. biloba 5.3 a 4.0 b 6 a 4 a 34 a 18 b 1.47 a 1.06 b
L. indica 13.3 a 11.3 a 39.A 28 b 1574 a 180 b 2.52 a 0.28 b
P. racemosa 40.7 a 27.3 b 673 a 287 b 2394 a 1561 b 0.26 a 0.38 a
Q. agrifolia 12.2 a 8.4 a 43 a 40 a 165 a 135 a 0.24 a 0.25 a
Q. lobata 12.5 a 13.8 a 91 a 66 b 664 a 316 b 0.49 a 0.34 b
S. sempervirens 7.3 a 4.0 a 64 a 15 b 404 a 37 b 0.36 a 0.03 b

Values followed by the same letter are not significantly different from one another at p=0.05 using Scheffe's multiple separation procedure.


Table 5. Total root length and total root area of trees growing with (+) and without (-) treeshelters after two years. Some roots <1 cm in diameter remained in the soil after root system excavation. (n=8 for 3- and 6-month data, n=20 for 1-year data, n=10 for 2-year data)

Species

Total Root Length (cm) Total Root Area (cm2)
- + - +
F. latifolia 1541 a 1099 b 1831 a 1034 b
G. biloba 273 a 166 b 138 a 52 b
L. indica 778 a 636 b 443 a 268 b
P. racemosa 2460 a 1713 b 2915 a 1809 b
Q. agrifolia 485 a 581 a 271 a 276 a
Q. lobata 730 a 605 b 555 a 353 b
S. sempervirens 692 a 539 b 1717 a 195 b

Values followed by the same letter are not significantly different from one another at p=0.05 using Scheffe's multiple separation procedure.


Table 6. Dry biomass response of Dendranthema grandiflora 'Bright Golden Anne' and Populus hybrida plants growing in aerated solution culture with (+) and without (-) treeshelters.

Dendranthema
Dry Weight (g) Root:Shoot

(dry weight basis)

Leaf Stem Root Total
- + - + - + - + - +
2.3 a 0.6 b 2.0 a 0.8 b 0.7 a 0.2 b 5.0 a 1.6 b 0.35 a 0.25 b
Populus
Dry Weight (g) Root:Shoot

(dry weight basis)

Leaf Stem Root Total
- + - + - + - + - +
37.6 a 2.8 b 15.7 a 3.3 b 11.5 a 1.5 b 64.8 a 7.6 b 0.73 a 0.45 b

Values followed by the same letter are not significantly different from one another at p=0.05 using Scheffe's multiple separation procedure.