<b>Fine-scale associational effects: single plant neighbours can alter susceptibility of focal plants to herbivores</b>

AbstractThe neighbourhood of plants in a patch can shape vulnerability of focal plants to herbivores, known as an associational effect. Associational effects of plant neighbourhoods are widely recognised. But whether a single neighbouring plant can exert an associational effect is unknown. Here, we tested if single neighbours indeed do influence the likelihood that a focal plant is visited and eaten by a mammalian herbivore. We then tested whether any refuge effect is strengthened by having more neighbours in direct proximity to a focal plant. We used native plant species and a browser/mixed feeder mammalian herbivore (swamp wallabies (<i>Wallabia bicolor</i>)) free-ranging in natural vegetation. We found that a single neighbouring plant did elicit associational effects. Specifically, plant pairs consisting of one high-quality seedling next to a single low-quality plant were visited and browsed by wallabies later and less than pairs of two high-quality seedlings. Having more neighbours did not strengthen these associational effects. Compared with no neighbours, one or five low-quality neighbours had the same effect in delaying time taken for wallabies to first visit a plot and browse on a high-quality focal seedling. While traditionally a 'patch' refers to a broad sphere-of-influence neighbouring plants have on a focal plant, our findings suggest the influence of plant neighbours can range from the nearest individual neighbour to the entire plant neighbourhood. Such fine-scale associational effects are fundamentally important for understanding intricate plant-herbivore interactions, and ecologically important by potentially having knock-on effects on plant survival, in turn influencing plant community structure.<b>Study site</b>Experiments were run in Ku- ring- gai Chase National Park, New South Wales, Australia (33°41′33″S, 151°08′44″E) from May 2021 to November 2022. All experimental sites were in <i>Eucalyptus</i> woodland dominated by scribbly gum (<i>Eucalyptus haemastoma</i>), red bloodwood (<i>Corymbia gummifera</i>), yellow bloodwood (<i>Corymbia eximia</i>), and grey gum (<i>Eucalyptus punctata</i>, our focal plant species)* *with complex shrub and ground vegetation layers. Animal ethics approval was granted by the University of Sydney’s Animal Ethics Committee (protocol number 2022/2196) and scientific license (SL102186) was obtained from New South Wales Office of Environment and Heritage to undertake the fieldwork.<b>Aim 1: Does a single plant neighbour elicit an associational effect?</b>Using these three species, we tested three paired treatments (Pair, 3 levels) using one plant of each of two species adjacent to one another, and one pair per plot. One treatment paired an <i>E. punctata</i> seedling with a <i>C. gummifera</i> seedling (n = 17 plots). Another paired an <i>E. punctata</i> seedling with a <i>B. pinnata</i> plant (n = 23 plots). The third treatment paired a <i>C. gummifera</i> seedling with a <i>B. pinnata</i> plant (n = 10 plots). The smaller sample size for the <i>C. gummifera</i> and <i>B. pinnata</i> combination was due to limited availability of <i>C. gummifera</i> seedlings. All three pairs were deployed in a completely randomised design at our study site (total n = 50 plots). We placed plots a minimum of 30 m apart, so one plot could not be seen from another. We used naturally occurring <i>B. pinnata</i> shrubs growing at the site (375 ± 19 mm). Seedlings of <i>E. punctata</i> and <i>C. gummifera</i> were of similar height (370 ± 25 mm), planted in soil in tube stock pots (10 cm diam), and supplied by one nursery (Plants Plus Cumberland Forest Nursery, West Pennant Hills, Sydney).For the treatment pairing <i>E. punctata</i> and <i>C. *gummifera,</i> <i>seedlings were placed 20 cm apart at either end of a clear tub (Fig 2a), allowing easy identification of which species was browsed in video footage For treatments involving *B. pinnata</i>, the eucalypt seedling was positioned at one end of a clear plastic tub, with the ‘free’ end of the tub placed next to a <i>B. pinnata</i> shrub, again 20 cm apart. A seedling pot with soil was included at the <i>B. pinnata</i> end of the tub to account for any effects of open soil (Fig 2b). The side on which each species was presented within each treatment was randomised by coin toss.At each Pair we set one motion-triggered infra-red trail camera (ScoutGuard SG560K or SG2060-K; Professional Trapping Supplies Pty Ltd, Molendinar, QLD, Australia) fastened to a wooden stake (camera height = 0.7 m) and placed directly over the pair of plants. Cameras were set to record for 1- minute with instant re-trigger. These videos were used to quantify the time to first visit a plot (days) by a wallaby (hereafter <i>time to 1st visit</i>), and time to first browse a plant (days) within a plot by a wallaby (hereafter <i>time to 1st browse</i>). If browsed, we visually estimated the percentage of foliage consumed (% intervals of 0, 25, 50, 75 and 100%) from each plant as seen on camera.<b>Aim 2: Does a neighbourhood of plants elicit stronger associational effects?</b>Here, we tested whether a refuge effect is strengthened by more low-quality neighbours in direct proximity to a focal plant, using the high-quality <i>E. punctata</i> and low-quality <i>B. pinnata</i> Pair. All plants, including <i>B. pinnata</i>, were in soil in pots (25 cm diam), and obtained from one nursery (Plants Plus Cumberland Forest Nursery, West Pennant Hills, Sydney).We compared four Treatments (<i>n</i> = 16 per Treatment); a single focal <i>E. punctata</i> seedling next to a single <i>B. pinnata</i> (single neighbour, Fig 3a), a single focal <i>E. punctata</i> surrounded by five B. pinnata* <i>(neighbourhood, Fig 3b), a single focal *E. punctata</i> with no neighbours as an untreated control** <i>*(Fig 3c), and a single focal *E. punctata</i> surrounded by five pots filled with soil as a procedural control (Fig 3d). All four Treatments were deployed at our study site in plots at least 30 m apart in a completely randomised design (total n = 64 plots). In the single neighbour treatment, <i>B. pinnata</i> was placed 50 cm away from the focal <i>E. punctata</i>. In the neighbourhood treatment, <i>B. pinnata</i> plants, and in the procedural control, pots with soil, were evenly spaced around the focal <i>E. punctata</i> at a 50 cm radius. This spacing ensured that <i>B. pinnata</i> did not completely visually obscure or physically obstruct wallabies’ access to the focal <i>E. punctata</i> seedlings.We calculated a visibility score for the single focal plant per plot, to be used in later analyses as a potential explanation of any observed associational refuge effects. Ordinal scores using visual estimates were used to determine how much of the focal <i>E. punctata *seedling could be seen from 1 m distance at an average adult swamp wallaby height of 70 cm [31]; 0 = 0%, 1 = 1– 25%, 2 = 26 – 50%, 3 = 51 – 75%, 4 = 76 – 100%; where percentages refer to the proportion of the focal *E. punctata</i> seedling visible. Visibility point scores were taken at 10 evenly spaced locations around each seedling. Point scores were later converted to midpoint values and an average used as a single visibility score.At each plot we set a motion-triggered infra-red trial camera (ScoutGuard SG560K or SG2060-K; Professional Trapping Supplies Pty Ltd, Molendinar, QLD, Australia). The camera was fastened to a wooden stake (camera height = 0.7 m, distance to seedling = 1.5 m) at an approximate 45° angle towards the plants (Fig 3c and d). Cameras were set to record for 1-minute with instant re-trigger. As per the single neighbour trial, these videos were used to quantify the <i>time to 1st visit</i> a plot (by a wallaby, in days), and <i>time to 1st browse *a focal *E. punctata *(by a wallaby, in days) within a plot. If browsed, we estimated the percentage of foliage consumed from the</i> E. punctata* seedling as for the single neighbour trial. We ran this experiment for 9 weeks, which was the time it took for at least 90% of plots of each Treatment to be visited by wallabies.<br><br>METADATA:<br><br>Files and variablesFile: Finnerty_etal_AIM1.csv<b>Description:</b> Data collected for Aim 1Variablesorder: To re-order CSVplotID: ID of the plot, i.e., ‘1PG’ = plot 1 of <i>E. punctata</i> and <i>C. gummifera</i> pairingtreatment: the pair type, i.e., PG = <i>E. punctata</i> and <i>C. gummifera</i> pairing, PB = <i>E. punctata</i> and <i>B. pinnata</i>, and GB = <i>C. gummifera</i> and <i>B. pinnata</i>Species: focal species within a pairNextTo: what species the focal species was next toSpeciesWithinPair: focal species within a pairFocalPlant: was this the focal species yes/noSetUpDateTime: set up time and dateVisted01: was the pair visited (0 = no, 1 = yes, NA = NA and see comment)dateVisit1: when the pair was visitedvisit1time: how long it took the pair to be visited (hours)visit1timeDAYS: how long it took the pair to be visited (days)visit1timeCen: how long it took the pair to be visited (hours - censored for survival analysis. I.e., not visited ‘0’ = 168 hours (until end of trial))visitCensoredData: was the visit data censored (Yes, No, NA)visit1Browse: was the pair browsed at first visit (1 = yes, 0 = no)Browsed01: was the pair browsed at all (including at first visit) (0 = no, 1 = yes, NA = NA and see comment)Bothspecnotbrow: where both species not browsed at all (0 = no, 1 = yes, NA = NA and see comment)SPECIESBrowsedate: when was the species within the pair was browsedSPECIESBrowseTime: how long it took for the species within the pair to be browsed (hours)SPECIESBrowseTimeCen: how long it took the species within the pair to be browsed (hours - censored for survival analysis. I.e., not visited ‘0’ = 168 hours (until end of trial))oneinpairbrow: was at least species within the pair browsed (1 = yes, 0 = no)BrowseCensoredData: was the browse data censored (Yes, No, NA)pBrowse: the proportion of the species within the pair browsed (0 = none browsed, 100 = whole plant browsed)BETA: pBrowsed turned into a true proportion for Beta analysisbrowseCompleteQ: data checked and correct (1 = yes)Comments: comments (i.e., why NA)File: Finnerty_etal_AIM1.R<b>Description:</b> R code analysis for Aim 1File: Finnerty_etal_AIM2_Midpoints.csv<b>Description:</b> Data collected for the focal species visability analysis of Aim 2VariablesTreatment: Treatment (as listed in the paper)Treat: Treatment ID shorthandTreatID: Treatment IDSD: SD card usedCamera: camera trap ID usedDateout: date cameras were deployedMidpointAV: % visability of focal species from 5 point analysis (see methods)File: Finnerty_etal_AIM2.csv<b>Description:</b> Data collected for Aim 2VariablesTreatment: Treatment (as listed in the paper)plotID: Plot ID (I.e., BH1 = <i>B. pinnata</i> neighbourhood replicate 1)SetUpDateTime: Date and time treatment was set upCameraNum: Camera trap number usedSDNum1: SD card number usedSDCheckDate: SD card check dateVisted01: was the treatment visited (0 = no, 1 = yes, NA = NA and see comment)dateVisit1: when the treatment was visitedvisit1timeDAYS: how long it took the treatment to be visited (days)visit1timeCenDAYS: how long it took the treatment to be visited (days - censored for survival analysis. I.e., not visited ‘0’ = 63 days (until end of trial))visitCensoredData: was the visit data censored (0 = no, 1 = yes, NA)visit1Browse: was the focal species browsed at first visit (1 = yes, 0 = no)Browsed01: was the focal species browsed at all (including at first visit) (0 = no, 1 = yes, NA = NA and see comment)dateVisitBrowse1: was the focal species browsed at first visit (1 = yes, 0 = no)dateBrowse1: when was the focal species browsedBrowseTimeDAYS: how long it took for the focal species to be browsed (days)BrowseTimeCenDAYS: how long it took the focal species to be browsed (days - censored for survival analysis. I.e., not visited ‘0’ = 63 hours (until end of trial))BrowseCensoredData: was the browse data censored (Yes, No, NA)EPBiounitsOut: how many ‘units’ of focal* E. puncata* where deployed within this replicate (see methods as to how a unit was defined)BioUnitsEatFirstBrowse: how many ‘units’ of focal* E. puncata* was browsed at first browsing visitppBrowse: how many ‘units’ of focal* E. puncata* browsed as a %BETA: how many ‘units’ of focal* E. puncata* browsed as a proportionNoVisorHab: data check (0 = all fine)Remove fomr analysis: 0 = check for analysis, 1 = remove from analysis (see comment as to why)Comments: comments (i.e., why NA)File: Finnerty_etal_AIM2.R<b>Description:</b> R code analysis for Aim 2Code/softwareMicrosoft Excel and R<br><br>