Supporting Online Material for

Size: px

Start display at page:

Download "Supporting Online Material for"

Victor Dawson
5 years ago
Views:

www.sciencemag.org/cgi/content/full/328/5986/1657/dc1 Supporting Online Material for Plants Integrate Information About Nutrients and Neighbors James F. Cahill Jr.,* Gordon G. McNickle, Joshua J.

1 Supporting Online Material for Plants Integrate Information About Nutrients and Neighbors James F. Cahill Jr.,* Gordon G. McNickle, Joshua J. Haag, Eric G. Lamb, Samson M. Nyanumba, Colleen C. St. Clair *To whom correspondence should be addressed. This PDF file includes: Materials and Methods Fig. S1 Tables S1 to S3 References Published 25 June 2010, Science 328, 1657 (2010) DOI: /science

2 Supporting Online Material Abstract Animals regularly integrate information about the location of resources and the presence of competitors, altering their foraging behavior accordingly. We studied the annual plant Abutilon theophrasti to determine whether a plant can demonstrate a similarly complex response to two conditions: presence of a competitor and heterogeneous resource distributions. Individually grown plants fully explored the pot using a broad and uniform rooting distribution regardless of soil resource distributions. Plants with competitors and uniform soil nutrient distributions exhibited pronounced reductions in rooting breadth and spatial soil segregation among the competing individuals. In contrast, plants with competitors and heterogeneous soil nutrient distributions reduced their root growth only modestly, indicating that plants synthesize information about both neighbor and resource distributions in determining their root behavior. Materials and Methods Study Species Abutilon theophrasti is an annual weed native to Asia and introduced into agricultural areas throughout the western hemisphere (S1). Prior studies indicate this species responds to soil heterogeneity through altered root growth (S2-4). Soil heterogeneity can cause increased plant biomass in individually grown plants (S4) and alter rooting breadth (S5), but has little impact on population characteristics (S2, S3). Whether this species demonstrates additive or situationspecific behavioral responses to variation in both resource distribution and the presence of competitors is unknown. Experimental Arenas In total there were six treatment combinations, representing all factorial combinations of soil heterogeneity (uniform, patch-center, patch-edge) and plant density treatments (alone vs. with competitor). We constructed plywood pots (mesocosms) that were 13.5cm long, 12.5 cm wide, and 25 cm deep. To record root growth over time, we placed a single clear acrylic tube horizontally through each mesocosm, 10 cm below the soil surface. One replicate of all six treatment combinations was placed onto a single tube, with a total of 10 replicate tubes used in the experiment. We were primarily interested in the interaction of roots between the focal and competitor plant. Therefore we limited our image analysis to only those locations under the focal plant, and in the direction towards the competitor plant. Thus, within each mesocosm we recorded 8 discrete, non-overlapping 13.5 x 18 mm image frames every two weeks for the duration of the experiment (from 2.7cm-13.5cm; Fig 1). Images were taken with a BTC 2 Minirhizotron camera system (Bartz Technology Corporation, Santa Barbara, CA, USA) at 15x magnification. During imaging, the camera was oriented upwards towards the soil surface, such that any roots which grew down would contact the tube and be visible to the camera. All mesocosms were filled with a background soil consisting of a low nutrient 3:1 mixture of sand and topsoil. Nutrient levels in all mesocosms were augmented by the addition of a small volume (25 cm 3 ) of a 1:1 mixture of sterilized steer manure and the background soil. The three soil treatments varied in the spatial distribution of these nutrients with either: (1) all 1

3 spread evenly throughout the soil (uniform), (2) one-half spread throughout the soil and one-half concentrated in a single rectangular column, 1.35 cm x 1.85 cm x 10 cm deep, located in the centre of the mesocosms (patch center), and (3) one-half spread throughout the soil and one-half concentrated in a single rectangular column located approximately 2 cm from the outside edge of the mesocosms (patch edge). Commercially available seeds of A. theophrasti were germinated in Petri dishes on moistened filter paper and bare-root transplanted into the pots after six days. All mesocosms had one focal individual planted approximately 3.4 cm from one edge (2.5 cm to the inside of the patch in both the patch-edge, and patch-center soil treatments; Fig. 1). In the alone treatment, the focal individual was the only plant growing in the mesocosms. In the with competition treatment, a second, equal-aged, individual was planted into the opposite side of the mesocosms (2.5 cm to the inside of the patch in the patch-center treatment, but 5 cm from the opposite edge; Fig. 1). The high nutrient patch in the patch-centre soil treatment was equidistant from the base of the two plants. The spacing of plants and patches was limited by the geometry of the minirhizotron imaging system, and we felt that it was more important to equally space the plants from each other and from the patch-center, than to equally space them from the walls of the mesocosm. We recognize that the walls of the mesocosm would influence root growth, and thus we did not analyze root growth in the direction of the nearest wall. Because all focal plants in all treatments were planted in the same location within the mesocosms, regardless of treatment, this planting design should not have caused the treatment effects observed. Growing conditions All mesocosms were placed in a growth room within the Biotron facilities of the Department of Biological Sciences at the University of Alberta. The room was set to a constant temperature of 22 o C, with a 16:8 L:D light cycle. Plants were watered regularly and allowed to grow for 8 weeks. Photosynthetically active radiation within the room was approximately 180 µmol/m 2 /sec. Plant Measurements, Root Imaging, and Root Staining Over the course of the experiment, we took digital images of the soil and roots in ten frames which formed a belt transect in each tube. Imaging began immediately following planting and then occurred every 2 weeks thereafter. After 8 weeks of growth, we took a final series of digital images, and then clipped all plants 3 cm from the soil surface. The results presented in this study come from this final round of images. Foraging models require understanding of individual movement and growth, and thus we modified a staining method (S6) to determine the identity of the roots in each frame in all mesocosms. We used a fast-curing epoxy putty compound to secure a short piece of ¼ clear plastic aquarium tubing over each stem, the other end of which had been previous glued to a 30 ml syringe. We placed 5 ml of a 1% solution of either acid fuschin or acid green dye into each tube, with each dye present only once in each mesocosms. The dye was placed under ~2 atm of pressure for 36 hours to allow sufficient time for penetration into the entire root system, after which time we took another round of root images. The digital images of the roots obtained before and after staining were analyzed using RooTracker 2.1 (Duke University, Durham, NC, USA). Presence or absence of roots at each location was recorded. We also recorded the 2

4 maximum rooting breadth by tracing a straight line from the base of the stem to the furthest visible edge of the root system. We also traced roots to estimate abundance. However, because root densities were generally low, measures of root abundance were highly variable, and their analysis without substantial meaning. We identified the roots to the target or competitor plant based on stain color and/or visible connection to other roots of known ownership. Across all treatments, 70% of the with competition pots had at least 90% of all root identified to individual. The overall patterns in the data did not vary as a function of whether we used all pots in the analyses, or only those for whom 90% root identity was achieved. From these data we were able to calculate the distribution of roots of each individual plant in relation to those of neighboring plants and the soil nutrient patches. Following staining, the remaining portion of the stem was clipped at the soil surface, and roots were extracted by submersion in water to facilitate easy separation of focal and competitor plant roots. Roots and shoots were dried at 70 C for 3 days and weighed. Statistical Analyses Soil occupancy. A generalized linear mixed model was fit using the SAS 9.2 GLIMMIX procedure (S7) to separate the effects of competition and soil heterogeneity on patterns of microhabitat use of the focal plant. In each image location, we recorded the presence or absence of roots of the focal plant. In the analyses, a binomial error distribution and logit link function were used. Competition (alone vs. with competition) and soil treatment (uniform, patch-center, patchedge) were included as fixed factors. Distance away from the focal plant (mm) and focal shoot biomass (natural log transformed) were included as continuous variables. The latter was included to control for expected decreases in soil occupancy far from the focal plant stem due simply to changes in plant size. All possible two and three-way interactions between distance, competition, and soil heterogeneity were included in the statistical model. Pot ID was included as a random factor to account for the multiple distance measurements taken in each pot. This analysis was repeated with and without plant mass as a covariate to attempt to control for allometric effects. The plant mass covariate was insignificant, and resulted in no material changes in the statistical results (Table S1). Thus, we can conclude that the shifts in rooting breadth and rooting distribution that we observed were not caused by allometric shifts in growth, but rather by behavioral changes in habitat use. Separate analyses investigating patterns of root co-occurrence among the focal plant and its competitor could not be conducted, as there were very few co-occurrences (15 out of a total of 210 possible frames), and model convergence could not be achieved. We interpret the low number of co-occurrences to indicate multiple individuals rarely co-occur at the micro-habitat scale. Foraging breadth. A general linear model was conducted to determine whether soil treatment and competition influenced the average maximum rooting breadth of the focal plants. Rooting breadth was determined as the greatest linear distance between the stem of the focal plant and the furthest root observed in the final round of images. The value served as the response variable, with soil treatment, competition treatment, and their interaction serving as 3

5 fixed effects. A priori comparison were conducted using LS means contrasts to test whether rooting breadth differed as a function of soil treatment when plants grew with a competitor. Plant biomass. Three general linear models were conducted to determine whether soil treatment and competition altered focal plant biomass. In each model, one aspect of plant biomass (shoot, root, total) served as the response variable, with soil treatment, competition treatment, and their interaction serving as fixed effects. In each model only competition was significant, and competition reduced the biomass of roots and shoots. 4

6 Figure S1 Plant biomass responses to soil and competition treatments. Presented below are focal plant shoot biomass (above the horizontal line) and root biomass (below the horizontal line) for individually grown plants (red) and plants grown with a competitor (blue). The x-axis corresponds to the three soil treatments that differed in the spatial arrangement of mineral resources. As indicated in Table S3, competition reduced plant biomass, while soil treatment had no effects on any component of plant biomass. 5

7 Table S1 Results from two generalized linear mixed models. The first model examines the influence of distance from the focal plant, competition, and soil treatment on the probability of the focal plant having any roots (binomial response variable) in that soil location. The second model is identical with the exception that focal shoot biomass was included as a covariate to test whether any observed differences in root distributions were due to changes in plant size. The significant three-way interactions indicate non-additive effects of competition and nutrient distributions of foraging behavior in this species. Without Covariate With Covariate Df F P Df F P Ln focal shoot biomass , Distance from focal plant 1, < , < Competition treatment 1, , Distance by competition 1, < , < Soil treatment 2, , Distance by soil treatment 2, , Competition by soil 2, , Distance by competition by soil 2, ,

8 Table S2 Results from a general linear model with rooting breadth of the focal plant as the normally distributed response variable. Rooting breadth was measured as the furthest distance from the plant stem a root of the focal plant was observed with the mini-rhizotron camera. Planned comparisons (LS mean contrasts) were conducted to determine statistical significance among the soil treatments when plants grew with a competitor. Df F P Competition treatment 1, <0.001 Soil treatment 2, Competition by soil 2, Planned comparisons (when grown with competition only) Uniform vs Patch-Center Uniform vs Patch-Edge Patch-Center vs Patch-Edge

9 Table S3 Results from three general linear models. In each model, one aspect of focal plant biomass (shoot, root, total) served as the response variable, with competition and soil treatments serving as fixed effects. Shoot Biomass DF F P Competition treatment 2, <0.001 Soil treatment 1, Competition by soil 2, Root Biomass DF F P Competition treatment 2, <0.001 Soil treatment 1, Competition by soil 2, Total Biomass Competition treatment 2, <0.001 Soil treatment 1, Competition by soil 2,

10 Supplemental references S1. N. R. Spencer, Economic Botany 38, 407 (1984). S2. B. B. Casper, J. F. Cahill, American Journal of Botany 83, 333 (1996). S3. B. B. Casper, J. F. Cahill, Jr., American Journal of Botany 85, 1680 (1998). S4. E. G. Lamb, J. J. Haag, J. F. Cahill, Functional Ecology 18, 836 (2004). S5. B. B. Casper, J. F. Cahill, R. B. Jackson, in The ecological consequences of environmental heterogeneity, M. J. Hutchings, E. A. John, A. J. A. Stewart, Eds. (Blackwell Science, Oxford, 2000), pp S6. C. Holzapfel, P. Alpert, Oecologia 134, 72 (2003). S7. SAS. (SAS Institute, Cary, NC, 2003). 9

Nursery experiments for improving plant quality

Nursery experiments for improving plant quality Why try nursery experiments? Often the common production techniques are used without experimenting with other procedures. Even if growth has been adequate