{"type": "FeatureCollection", "features": [{"id": "10.5061/dryad.51c59zwgj", "type": "Feature", "geometry": null, "properties": {"updated": "2026-09-22T16:19:42Z", "type": "Dataset", "created": "2024-04-02", "title": "Data from: Evidence for reductions in physical and chemical plant defense traits in island flora", "description": "Open Access# Evidence for Reductions in Physical and Chemical Plant Defense Traits in Island Flora [https://doi.org/10.5061/dryad.51c59zwgj](https://doi.org/10.5061/dryad.51c59zwgj) This dataset consists of three primary data sources: (1) Morphological and chemical measurements of leaf traits, collected from five taxonomic pairs of chaparral shrubs (*Ceanothus megacarpus*, *Cercocarpus betuloides*, *Dendromecon rigida/harfordii*, *Heteromeles arbutifolia*, *Prunus ilicifolia*) at three sites on the California Channel Islands (Santa Rosa, Santa Cruz, Santa Catalina) and three sites on the California mainland. (2) Morphological and chemical measurements of the same leaf traits from the same species, but this time measured from plants growing at botanic gardens (3) Morphological, chemical, and biomass data from a common garden experiment with *Stachys bullata*, with genotypes from two islands (Santa Rosa, Santa Cruz) and four mainland locations In addition, our analysis also includes bioclimatic data and local precipitation data accessed from publicly available sources. ## Description of the data and file structure This dataset is organized into two folders: **data_files** and **scripts** --- ***DATA_FILES*** Within the **data_files** folder, there are folders for '**Shrubs**' (corresponding to 1 and 2 above) and '**Stachys**' (corresponding to 3 above). **SHRUBS** The **Shrubs** folder contains one file (**Bowen and Van Vuren Effect Sizes.xlsx**), which summarizes the results from Bowen and Van Vuren (1997 ([https://www.jstor.org/stable/2387407](https://www.jstor.org/stable/2387407), directly as reported in their Tables 2, 3, 4, and 5 in the main text. Variables in this datafile include: 1. Trait - the plant trait that was measured in their study 2. Genus - the taxonomic unit being measured 3. t - the value of the t-statistic from a paired t-test of island vs. mainland samples for a given genus 4. n island - sample size for island plants 5. n mainland - sample size for mainland plants 6. Cohen's D - derived value that expresses insularity effect size for a given measure The **Shrubs** folder also contains four subfolders: **Cyanide**, **Images**, **Mapping**, and **Morphology** The **Cyanide** folder contains two files: 1. **cyanide_calibration.csv** - file containing measurements used to define calibration curve for quantifying evolved HCN from leaf tissue. 1. conc = concentration of potassium cyanide (KCN) standard used in calibration (mg/L) 2. abs = absorbance value returned by VWR V-1200 spectrometer, measured at 510 nm 2. **cyanide_measurements.csv** - file containing measurements of evolved HCN from field and botanic garden leaf tissue. PlantID values are the same as those reported for all other morphological measurements. 'NA' values in this dataset correspond to samples whose absorbance values were outside the range of our calibration curve or that were otherwise not suitable to include in analysis. 1. Age = whether leaf tissue was newly expanded ('young') or mature ('old') 2. Tissue_Mass = amount of frozen tissue used in assay (mg) 3. Dilution 1 = amount of water (mL) into which evolved HCN (in NaOH) was added prior to titration with citric acid. This value is 30 mL for all samples. 4. Dilution 2 = dilution factor. Here, a value of 1 means that 5 mL of citrate buffer was mixed with 5 mL water (1:1 ratio) and used in the subsequent reaction. A value of 10 means that 1 mL of citrate buffer was mixed with 10 mL water (1:10 ratio). 5. Sample Concentration = concentration of HCN in sample (mg/L), calculated using the calibration curve above. Samples with absorbance values above 0.500 were omitted and re-measured at reduced concentration, as this was beyond the concentration limit recommended by the manufacturer instructions. 6. Tissue Concentration = value relating dilution factor and sample mass to sample concentration. Expressed in milligrams of HCN per gram of leaf tissue. The **Images** folder contains all scanned leaf images (n = 626). File names correspond to plant species, plant ID, sampling site, and canopy position (see chaparral_leaf_morphology.csv below for a full description). So, for example, CMEG44_SMM_Upper refers to Ceanothus megacarpus, Plant ID = 44, sampled from the Santa Monica Mountains (SMM), upper canopy. Note also that each leaf within each image is individually numbered. The **Mapping** folder contains two files: 1. **shrubs_coordinates.csv** - contains coordinates and elevation for all field-sampled plants, recorded using a handheld Garmin GPS unit 2. **site_coordinates.csv** - contains broad site-level coordinates used for making map in Figure 1 The **Morphology** folder contains two files: 1. **chaparral_leaf_morphology.csv** - the primary datafile for this study, with each row (n = 5665) corresponding to a single leaf. For a visual depiction of the measurement protocol, see Supplemental Figures. Leaf measurements reported as NA generally correspond to leaves that were severely damaged, from which measurements could be reliably taken. 1. Index = sorting variable 2. IM = refers to whether a given plant was growing at an island or mainland site 3. Source = the original provenance of a given plant. For all field-sampled plants, the value here is the same as the value for 'Site' 4. Site = the location where plants were sampled. Includes all field sampling locations as well as the two botanic gardens 5. Exclosure = yes/no variable, only relevant to Catalina Island, describing whether sampled plant was inside of a deer exclosure 6. Species = taxon being measured 7. Plant = Plant ID, a unique value for each individual plant. Note that botanic garden samples have their own non-integer codes, and for Rancho Santa Ana Botanic Garden, these codes can be cross-referenced against the garden's living collections 8. Position = refers to whether a sampled branch came from the upper (>2m) or lower portion of the plant's canopy 9. Aspect = recorded from the Garmin GPS, refers to predominant downward slope direction. Not recorded for botanic garden plants (marked as NA) or for plants from completely flat ground. 10. Elevation = elevation in meters of sampled plants 11. Diameter1 = diameter (cm) of the primary plant trunk at 0.25m (NA means that stem could not be reliably measured) 12. Diameter2 = diameter (cm) of any secondary plant trunk at 0.25m (only applicable for multi-stemmed plants; NA means that stem could not be reliably measured) 13. Stem_Area = derived measure of stem area (cm^2), based on trunk diameter, used as a rough proxy for plant age (NA means that stem could not be reliably measured) 14. 1st_year = refers to whether an individual leaf was newly emerged growth (1) or fully expanded and mature (0) 15. Leaf_ID = corresponds to the numbers in each leaf scan; identifies each individual leaf from a given branch 16. Leaf_Length = leaf length (cm) along its primary axis, excluding the petiole 17. Leaf_Area_petiole = leaf area (cm^2), including the petiole 18. Leaf_Area_no.petiole = leaf area (cm^2), excluding the petiole 19. Internal_area_correction = cumulative area of any 'holes' missing within the leaf perimeter (cm^2) 20. True_area = Leaf_Area_no.petiole minus Internal_area_correction (cm^2) 21. Leaf_area_corrected = leaf area, after manually filling in gaps missing due to presumed herbivore damage (cm^2) 22. Leaf_area_corrected_final = Leaf_area_corrected minus Internal_area_correction (cm^2) 23. Area_no_spines = leaf area after connecting vertices created by leaf spines (cm^2), using to calculate spinescence (%) 2. **shrub_leaf_masses.csv** - cumulative mass (g) of fully expanded leaf tissue from each branch, summed across all individual leaves. Used for calculating specific leaf area (SLA). **STACHYS** The **Stachys** folder contains three subfolders: **Chemistry**, **Morphology**, and **Setup** The **Chemistry** folder contains two files and one sub-directory: 1. **stachys_chromatograms** contains raw GC-MS readout for six leaf chemistry samples. Within each of the corresponding subfolders, the tic_front.csv file was used to generate the chromatograms shown in Figure 6A. 2. **stachys_compound_list.csv** is the full list of compounds detected in our samples. RT refers to the retention time (in minutes) of each compound. Identifications are putative. 3. **stachys_leaf_vocs.csv** is the full data matrix of leaf volatile compounds, with each sample as its own row and data columns each corresponding to a single compound. Values in this data matrix correspond to integrated peak areas, which are a proxy for the abundance of each compound. The **Morphology** folder contains two files: 1. **Anet-stbu.xlsx** contains gas exchange measurements for 26 plants measured in the common garden. The gas exchange column is net carbon assimilation, expressed as CO2 uptake per unit time per unit leaf area (\u00b5mol of CO2 m-2 s-1). 2. **sla_sbbg.csv** contains specific leaf area measurements for *Stachys* plants in the common garden. Note that plant #54 had died by the time of data collection, hence its values of NA across all columns. 1. ID = individual plant ID 2. SLA = cumulative area/ cumulative mass (cm^2/g) 3. leaves = refers to the number of leaves used for generating SLA measurement 4. area/leaf = cumulative area/ leaf number (cm^2/leaf) The **Setup** folder contains three files: 1. **321dailys.xls** is a file containing annual precipitation records (inches) for the Santa Barbara Botanic Garden, accessed from: [https://www.countyofsb.org/2328/Daily-Rainfall-Data-XLS](https://www.countyofsb.org/2328/Daily-Rainfall-Data-XLS) 2. **Field_Setup_SBBG.csv** is the primary file containing details on the primary garden experiment. Note that samples with masses recorded as NA were either dead at the time of sampling. Plants grown on Santa Cruz Island have values of NA for row and column, as this common garden was not arranged in a grid. 1. Index = individual plant ID 2. Population = provenance of plant 3. Garden = whether plants were grown at the Santa Barbara Botanic Garden (primary common garden site) or at the field station on Santa Cruz Island (secondary garden location with only Santa Cruz genotypes) 4. Genotype = identifier given to field-collected rhizomes, which were then propagated and split prior to planting out 5. Cumulative_Mass = mass (g) of paper bag and all of its contents, used for measuring end-of-season plant aboveground biomass 6. Bag_Mass = mass (g) of bag itself (without its contents) 7. Inside_Bag_Mass = mass (g) of smaller paper bags contained within larger bags, including all of their contents. Though not analyzed, these inside bags included all plant biomass collected from outside of the gopher cage that plants were growing in. 8. Inside_Bag_Only_Mass = as above, mass (g) of inner bag itself (without its contents) 9. Year = whether biomass was collected in 2016 or 2017 10. Row = grid location within common garden. Row 1 was at the bottom of the slope shown in Figure 2. 11. Column = grid location within common garden. 3. **stachys_coordinates.csv** contains coordinates for the six collecting sites, used to make the map in Figure 2. --- ***SCRIPTS*** All analyses for this project were conducted in the R programming language (version 4.1.3). Scripts used for analysis are arranged in two folders: **Shrubs** and **Stachys** The **Shrubs** folder contains the following scripts: 1. **coordinates_shrubs_stachys.R** - script used for generating all maps, including those in Figures 1 and 2 and the Google Earth maps in the supplementary figures 2. **cyanide_calibration.R** - script for plotting the calibration curve for relating evolved absorbance values to evolved HCN 3. **shrub_leaf_morphology_chemistry.R** - primary analysis script for manuscript, containing all major statistical analyses and plotting 4. **shrubs_BioClim.R** - script used for extracting bioclimatic data for field-sampled plants; containing code generating climate figures shown in supplementary materials The **Stachys** folder contains the following scripts: 1. **sbbg_precip_data.R** - very short script for summarizing water year totals for 2017 at the Santa Barbara Botanic Garden 2. **stachys_analysis.R** - primary script for generating all analyses and figures for *Stachys* common garden data 3. **stbu_gas_exchange.R** - script for analyzing gas exchange in common garden *Stachys* Note that for recreating some analyses and figures, users will need a Google Maps API key and will need to download data from the bioclim database. --- ## Sharing/Access information Data, code, and figures associated with this project are also available on GitHub at the following link: [https://github.com/micahfreedman/manuscripts/tree/master/Island_Mainland](https://github.com/micahfreedman/manuscripts/tree/master/Island_Mainland)", "keywords": ["Islands", "Morphology", "Dendromecon", "cyanogenic glycosides", "Ecology", "Terpenes", "Cercocarpus", "California Channel Islands", "Chemical ecology", "marginal spines", "Specific leaf area", "Plant science", "Heteromeles", "FOS: Biological sciences", "Stachys", "Other", "Prunus", "Herbivory", "Plant defenses", "Plant-herbivore interactions", "Ceanothus", "Ecology", " Evolution", " Behavior and Systematics"], "contacts": [{"organization": "Freedman, Micah", "roles": ["creator"]}]}, "links": [{"href": "https://doi.org/10.5061/dryad.51c59zwgj"}, {"rel": "self", "type": "application/geo+json", "title": "10.5061/dryad.51c59zwgj", "name": "item", "description": "10.5061/dryad.51c59zwgj", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main/items/10.5061/dryad.51c59zwgj"}, {"rel": "collection", "type": "application/json", "title": "Collection", "name": "collection", "description": "Collection", "href": "https://repository.soilwise-he.eu/cat/collections/metadata:main"}], "time": {"date": "2024-01-01T00:00:00Z"}}, {"id": "10.1002/ece3.6803", "type": "Feature", "geometry": null, "properties": {"updated": "2026-09-22T16:14:25Z", "type": "Journal Article", "created": "2020-09-22", "title": "Background insect herbivory increases with local elevation but makes minor contribution to element cycling along natural gradients in the Subarctic", "description": "Abstract

Herbivores can exert major controls over biogeochemical cycling. As invertebrates are highly sensitive to temperature shifts (ectothermal), the abundances of insects in high\uffe2\uff80\uff90latitude systems, where climate warming is rapid, is expected to increase. In subarctic mountain birch forests, research has focussed on geometrid moth outbreaks, while the contribution of background insect herbivory (BIH) to elemental cycling is poorly constrained. In northern Sweden, we estimated BIH along 9 elevational gradients distributed across a gradient in regional elevation, temperature, and precipitation to allow evaluation of consistency in local versus regional variation. We converted foliar loss via BIH to fluxes of C, nitrogen (N), and phosphorus (P) from the birch canopy to the soil to compare with other relevant soil inputs of the same elements and assessed different abiotic and biotic drivers of the observed variability. We found that leaf area loss due to BIH was ~1.6% on average. This is comparable to estimates from tundra, but considerably lower than ecosystems at lower latitudes. The C, N, and P fluxes from canopy to soil associated with BIH were 1\uffe2\uff80\uff932 orders of magnitude lower than the soil input from senesced litter and external nutrient sources such as biological N fixation, atmospheric deposition of N, and P weathering estimated from the literature. Despite the minor contribution to overall elemental cycling in subarctic birch forests, the higher quality and earlier timing of the input of herbivore deposits to soils compared to senesced litter may make this contribution disproportionally important for various ecosystem functions. BIH increased significantly with leaf N content as well as local elevation along each transect, yet showed no significant relationship with temperature or humidity, nor the commonly used temperature proxy, absolute elevation. The lack of consistency between the local and regional elevational trends calls for caution when using elevation gradients as climate proxies.Few studies have tested how plant quality and the presence of competitors interact in determining how herbivores choose between different leaves within a plant. We investigated this in two herbivorous spider mites sharing tomato plants: Tetranychus urticae, which generally induces plant defenses, and Tetranychus evansi, which suppresses them, creating asymmetrical effects on coinfesting competitors. On uninfested plants, both herbivore species preferred young leaves, coinciding with increased mite performance. On plants with heterospecifics, the mites did not prefer leaves on which they had a better performance. In particular, T.\uffc2\uffa0urticae avoided leaves infested with T.\uffc2\uffa0evansi, which is in agreement with T.\uffc2\uffa0urticae being outcompeted by T.\uffc2\uffa0evansi. In contrast, T.\uffc2\uffa0evansi did not avoid leaves with the other species, but distributed itself evenly over plants infested with heterospecifics. We hypothesize that this behavior of T.\uffc2\uffa0evansi may prevent further spread of T.\uffc2\uffa0urticae over the shared plant. Our results indicate that leaf age determines within\uffe2\uff80\uff90plant distribution of herbivores only in absence of competitors. Moreover, they show that this distribution depends on the order of arrival of competitors and on their effects on each other, with herbivores showing differences in behavior within the plant as a possible response to the outcome of those interactions.Few studies have tested how plant quality and the presence of competitors interact in determining how herbivores choose between different leaves within a plant. We investigated this in two herbivorous spider mites sharing tomato plants: Tetranychus urticae, which generally induces plant defenses, and Tetranychus evansi, which suppresses them, creating asymmetrical effects on coinfesting competitors. On uninfested plants, both herbivore species preferred young leaves, coinciding with increased mite performance. On plants with heterospecifics, the mites did not prefer leaves on which they had a better performance. In particular, T.\uffc2\uffa0urticae avoided leaves infested with T.\uffc2\uffa0evansi, which is in agreement with T.\uffc2\uffa0urticae being outcompeted by T.\uffc2\uffa0evansi. In contrast, T.\uffc2\uffa0evansi did not avoid leaves with the other species, but distributed itself evenly over plants infested with heterospecifics. We hypothesize that this behavior of T.\uffc2\uffa0evansi may prevent further spread of T.\uffc2\uffa0urticae over the shared plant. Our results indicate that leaf age determines within\uffe2\uff80\uff90plant distribution of herbivores only in absence of competitors. Moreover, they show that this distribution depends on the order of arrival of competitors and on their effects on each other, with herbivores showing differences in behavior within the plant as a possible response to the outcome of those interactions.Herbivores can exert major controls over biogeochemical cycling. As invertebrates are highly sensitive to temperature shifts (ectothermal), the abundances of insects in high\uffe2\uff80\uff90latitude systems, where climate warming is rapid, is expected to increase. In subarctic mountain birch forests, research has focussed on geometrid moth outbreaks, while the contribution of background insect herbivory (BIH) to elemental cycling is poorly constrained. In northern Sweden, we estimated BIH along 9 elevational gradients distributed across a gradient in regional elevation, temperature, and precipitation to allow evaluation of consistency in local versus regional variation. We converted foliar loss via BIH to fluxes of C, nitrogen (N), and phosphorus (P) from the birch canopy to the soil to compare with other relevant soil inputs of the same elements and assessed different abiotic and biotic drivers of the observed variability. We found that leaf area loss due to BIH was ~1.6% on average. This is comparable to estimates from tundra, but considerably lower than ecosystems at lower latitudes. The C, N, and P fluxes from canopy to soil associated with BIH were 1\uffe2\uff80\uff932 orders of magnitude lower than the soil input from senesced litter and external nutrient sources such as biological N fixation, atmospheric deposition of N, and P weathering estimated from the literature. Despite the minor contribution to overall elemental cycling in subarctic birch forests, the higher quality and earlier timing of the input of herbivore deposits to soils compared to senesced litter may make this contribution disproportionally important for various ecosystem functions. BIH increased significantly with leaf N content as well as local elevation along each transect, yet showed no significant relationship with temperature or humidity, nor the commonly used temperature proxy, absolute elevation. The lack of consistency between the local and regional elevational trends calls for caution when using elevation gradients as climate proxies.Herbivores can exert major controls over biogeochemical cycling. As invertebrates are highly sensitive to temperature shifts (ectothermal), the abundances of insects in high\uffe2\uff80\uff90latitude systems, where climate warming is rapid, is expected to increase. In subarctic mountain birch forests, research has focussed on geometrid moth outbreaks, while the contribution of background insect herbivory (BIH) to elemental cycling is poorly constrained. In northern Sweden, we estimated BIH along 9 elevational gradients distributed across a gradient in regional elevation, temperature, and precipitation to allow evaluation of consistency in local versus regional variation. We converted foliar loss via BIH to fluxes of C, nitrogen (N), and phosphorus (P) from the birch canopy to the soil to compare with other relevant soil inputs of the same elements and assessed different abiotic and biotic drivers of the observed variability. We found that leaf area loss due to BIH was ~1.6% on average. This is comparable to estimates from tundra, but considerably lower than ecosystems at lower latitudes. The C, N, and P fluxes from canopy to soil associated with BIH were 1\uffe2\uff80\uff932 orders of magnitude lower than the soil input from senesced litter and external nutrient sources such as biological N fixation, atmospheric deposition of N, and P weathering estimated from the literature. Despite the minor contribution to overall elemental cycling in subarctic birch forests, the higher quality and earlier timing of the input of herbivore deposits to soils compared to senesced litter may make this contribution disproportionally important for various ecosystem functions. BIH increased significantly with leaf N content as well as local elevation along each transect, yet showed no significant relationship with temperature or humidity, nor the commonly used temperature proxy, absolute elevation. The lack of consistency between the local and regional elevational trends calls for caution when using elevation gradients as climate proxies.