Data from: Predicting success of conservation translocations: Prerelease screening tools for a threatened marsupial
Abstract
Individual animals differ in their responses to environmental challenges. Identifying these differences before animals are translocated could be a useful tool in improving postrelease success. For most species, however, uncertainty persists around which behaviours contribute to postrelease success. Greater bilbies, Macrotis lagotis, are a threatened species, for which detailed behavioural studies in situ are rare. We tested bilbies prior to their release to a fenced safe haven to examine whether (1) bilby response to a known threat (that is, human observer) and novelty (that is, novel scent) was individually different and repeatable, or alternatively plastic, and (2) if any behaviours were related to bilby fitness (breeding success and survival) postrelease. We found that several behaviours measured during trapping (exposure to a known threat) were individually variable and repeatable and that males that were less responsive (lower respiratory rate) during handling had greater breeding success (that is, the number of offspring sired relative to the opportunity to breed). Contrastingly, individual differences in bilby response to novel scents were not repeatable, but we did, however, find sex-specific responses in this context. Males took less time to initiate touch with the novel-scented object and touched the object more times than females, suggesting that males are less neophobic. Plasticity rather than personality predicted female post-release fitness. Females that were increasingly cautious of novelty had greater relative breeding success. Our study is the first to confirm repeatable and plastic behaviours in bilbies, develop a standardized test for prerelease screening of behaviours, and examine the fitness consequences of these behaviours. We encourage the use of the ‘trap test’ as a simple and practical way to screen bilbies for future conservation translocations. We also recommend further testing in environments where the risk of establishment failure may be higher (that is, invasive predators present) and where the fitness consequences may be more profound and thus detectable.
https://doi.org/10.5061/dryad.95x69p8tg
This README file lists the supplementary datasets and R code used in the associated paper. It also provides a brief description of how each file was used in generating the results, and definitions for any abbreviated variables within datasets. Note: dates are all in the format ‘date-month-year’.
Description of the data and file structure
Data.zip
trap-test.csv
Behavioural observations collected during the trap test. Data includes individual identity (‘ind id’), sex (male = M, female = F), the release group (release gp) each individual belongs to (1 = first release, 2 = second release, 3 = third release), date the trap test was conducted (day/month/year), the order that each test was conducted (‘trial id’) with ‘1’ indicating the first test for the corresponding individual with following tests ascending in 1 unit increments, the length of time from first encountering a captured individual within a trap to the time the animal is released in minutes (handling time), whether the individual was manually restrained and removed from the trap (manual.removal = 1) or voluntarily entered the capture bag after the trap door was opened (manual.removal = 0), whether the individual moved backwards or forwards in the trap (move in trap = 1) or remained still in the trap (move in trap = 0), whether the individual moved when suspended (move when suspend = 1) or remained still (move when suspend = 0), whether the individual made a noise when handled (any noise = 1) or was quiet (any noise = 0), whether the individual remained in the open (direction of escape = 1) or went down into a burrow after release (direction of escape = 0), how long in seconds that it took an individual to enter a capture bag from a trap (latency to enter bag) to a maximum of 60s, the respiratory rate (breaths/minute) for each individual during handling (resp rate), how long in seconds it took an individual to leave the capture bag upon release (latency to leave), and the total amount of time in seconds that it took for an individual to escape out of sight (time to OOS) to a maximum of 120 seconds. More detail on how each of the behavioural responses were measured is included in the associated paper. Where a behavioural response was not measured for a particular trial id it is assigned a 'NA'.
Novel_scent_test.csv
Behavioural observations collected during the novel-scent test. Data includes individual identity (ind id), sex (male = M, female = F), the order that each test was conducted (‘trial id’) with ‘1’ indicating the first test for the corresponding individual with following tests ascending in 1 unit increments, the release group (release gp) each individual belongs to (1 = first release, 2 = second release, 3 = third release), date the novel-scent test was conducted (day/month/year), the time both the ears and head of each individual was observed outside the burrow entrance (time emerged), the scent used for the test (either SO = sardine oil, PB = peanut butter, H = honey), the number of times each individual touched the novel-scented object (No of touches), the difference in time in seconds between time emerged and time each individual first touched the novel-scented object (TTT) up to a maximum of 600 seconds, the difference in time in seconds between the time each individual first touched the novel-scented object and the first time an individual paused or moved away from the novel-scented object (TTS) up to a maximum of 600 seconds. More detail on how each of the behavioural responses were measured is included in the associated paper. The origin of the data, whether it was taken from in-person observations (IP) or from the video footage (F) or both (IP/F) are provided along with some general comments on the observations taken (comments). Where a behavioural response was not measured for a particular trial id it is assigned a 'NA'.
fitness_data.csv
Data collected for modelling the relationship between breeding success, survival and repeatable (personality) and non-repeatable (plastic) behaviours. Individual identity (ind id), sex (male = M, female = F), physical attributes (head length (mm), right-hind short pes (R short pes, mm), release weight (kg), and body condition score at release/release bcs (1-5)), the release group (release gp) each individual belongs to (1 = first release, 2 = second release, 3 = third release), the number of individuals that were held in the pre-release area together (no in pen), whether or not each of the (N = 19) bilbies released were confirmed alive (mortality = 0) or deceased (mortality = 1) after 4-months post-release, the number of resultant offspring of all (N = 18) mature bilby founders released (no.offspring), the estimated number of days each individual had to breed (estimated.days.obs), the order that each test was conducted (‘trial id’) with ‘1’ indicating the first test for the corresponding individual with following tests ascending in 1 unit increments, and each of the behavioural responses for only the behaviours that were individually different. Units are as described above for 'trap-test.csv' and 'Novel_scent_test.csv'. The bilby which was immature at release (ind id = 15) was not included in the breeding success analysis and so no offspring = NA and estimated days observed (estimated.days.obs) = NA. Where a behavioural response was not measured for a particular trial id it is assigned a 'NA'.
plasticity-fitness.csv
Data for modelling the relationship between breeding success, survival and non-repeatable (plastic) behaviours. Individual identity (ind id), sex (male = M, female = F), release group (release gp) that the individual belonged to (1 = first release, 2 = second release, 3 = third release), the number of resultant offspring (no.offspring) of all (N = 18) mature bilby founders released, whether or not each of the (N = 19) bilbies released were confirmed alive (mortality = 0) or deceased (mortality = 1) after 4-months post-release, the estimated number of days each individual had to breed (estimated.days.obs), and the slope coefficient extracted from the linear model for number of touches of the novel-scent by trial number for each individual (slope NOT). Individuals that have a slope NOT = NA had only one observation for the number of touches and so we could not calculate the slope for these individuals. The bilby which was immature at release (ind id = 15) was not included in the breeding success analysis and so no offspring = NA and estimated days observed (estimated.days.obs) = NA.
Bilby_parentage_and_relatedness.csv
Results of the parentage and relatedness analysis, including a unique index for each record (Index), log likelihood ratios of dam (LLRdam), sire (LLRsire), and parent pair (LLRpair) relationships assigned via ‘sequoia’, and the Triodic Maximum Likelihood (TrioML R) estimated pairwise relatedness and the lower (TrioML R 95LCI) and upper (TrioML R 95UCI) 95% confidence intervals for these estimates based on 1,000 bootstraps of the data for each bilby dyad indicated under the ‘Parent’ and ‘Offspring’ (PO) columns. The sex of each bilby for each dyad is also listed (Sexpair) for each combination (F-M- = female-male dyad, M-M- = male-male dyad, and F-F- = female-female dyad). Bilby dyad labels match the labels on the raw SNP data*. Delta8 and Delta7 are additional metrics that help further confirm the correct PO assignment as their values typify certain relationships (see Table 1). The 95% confidence intervals for the Delta8 (Delta8 95LCI, Delta8 95UCI) and Delta7 (Delta7 LCI, Delta7 95U) are also given. Some general comments are included (Comments). LLRdam = NA when the parent assigned is the father (i.e. no dam), and LLRsire = NA when the parent assigned is the mother (i.e. no sire). LLRpair = NA when one of the parents are unsampled and the pair of parents cannot be assigned.
Table 1: Relatedness and identical by descent (delta 7 = ∆7 and delta 8 = ∆8) probabilities and the typical relationship associated with each probability. Probabilities are obtained from Wang 2011.
| Relationship | Relatedness | ∆7 | ∆8 |
|---|---|---|---|
| Parent-offspring | 0.5 | 0 | 1 |
| Full sibs | 0.5 | 0.25 | 0.5 |
| Half sibs | 0.25 | 0 | 0.5 |
| Grandparent-grandchild | 0.25 | 0 | 0.5 |
| Avuncular (i.e. any combination of uncle, aunty, niece and nephew) | 0.25 | 0 | 0.5 |
Code/Software
R code.zip
Bilby pre-release screening analysis.Rmd
Annotated R code to replicate all analyses, plots, and modelling as described in the associated paper. This includes annotated R code for filtering and formatting raw SNP data (from DArTSeq) into a form that COANCESTRY can work with. Output of this code is saved and inputted into the separate COANCESTRY software. Code for formatting was shared and written by Katherine Farquharson. For the code to run the raw SNP data (from DArTSeq) needs to be in your working directory
Raw SNP data from the associated paper (part of other unpublished manuscript which is under review process) can be requested from the corresponding author (Kate Cornelsen: cornelsen.kate@gmail.com) upon reasonable request.
Data includes the behavioural observations taken during each personality test (trap and novel scent), and the survival and breeding success statistics of greater bilby (Macrotis lagotis) founders included in the study. Further details on the methods used to collect data are described in the associated paper. We have also provided the R code used to assess between-individual behavioural differences (i.e., personality), consistency of within-individual responses to personality tests (i.e., repeatability), and for assessing the relationship between bilby personality and plastic behaviours and post-release fitness (i.e., survival and breeding success). Additionally, R code to replicate the parentage analysis is also provided along with the results of the analysis.