Measuring ontogenetic shifts in central-place foraging insects: a case study with honey bees

Measuring time-activity budgets over the complete individual lifespan is now possible for many animals with the recent advances of life-long individual monitoring devices. Although analyses of changes in the patterns of time-activity budgets have revealed ontogenetic shifts in birds or mammals, no such technique has been applied to date on insects. We tested an automated breakpoint-based procedure to detect, assess and quantify shifts in the temporal pattern of the flight activities in honey bees. We assumed that the learning and foraging stages of honey bees will differ in several respects, to detect the age at onset of foraging (AOF). Using an extensive dataset covering the life-long monitoring of 2,100 individuals, we compared the AOF outputs with the more conventional approaches based on arbitrary thresholds. We further evaluated the robustness of the different methods comparing the foraging time-activity budget allocations between the presumed foragers and confirmed foragers. We revealed a clear-cut learning-foraging ontogenetic shift that differs in duration, frequency, and time of occurrence of flights. Although AOF appeared to be highly plastic among bees, the breakpoint-based procedure seems better able to detect it than arbitrary threshold-based methods that are unable to deal with inter-individual variation. We developed the aof R-package including a broad range of examples with both simulated and empirical dataset to illustrate the simplicity of use of the procedure. This simple procedure is generic enough to be derived from any individual life-long monitoring devices recording the time-activity budgets of honey bees, and could propose new ecological applications of bio-logging to detect ontogenetic shifts in the behaviour of central-place foraging insects.

of 2,100 individual honey bees. We further compared our AOF outputs with the more 159 conventional approaches based on arbitrary thresholds, evaluated the robustness of 160 the procedure with different types of datasets, and compared the foraging time-161 activity budget allocations between the predicted foragers and observed foragers. 162 We also explored the relationship between AOF ontogenetic plasticity and foraging 163 performance in honey bees, and we tracked the potential risk of early ontogenetic  Individual honey bees were monitored using RFID tags (mic3® -TAG 64 bit RO, 186 iID2000, 13.56 MHz system, 1.0 mm × 1.6 mm × 0.5 mm; Microsensys GmbH, Erfurt, 187 Germany) and ad-hoc readers (iID2000, 2k6 HEAD; Microsensys GmbH, Erfurt,

188
Germany) placed at the entrance of the hives (Figure 1a). Readers were powered by 189 a turnover of two slow-discharge rate batteries (12V, 92Ah, C20, Banner®). The     (Table S1). Each honey bee was 211 equipped with a unique RFID tag, and released in cohorts of 75-150 individuals into 212 the three experimental colonies (Table S1).

213
The time-activity budgets were recorded using two adjacent rows of five 214 contiguous RFID readers placed at the entrance of the hives (i.e. the common 10-  (https://cran.r-project.org/package=aof) and GitHub (https://github.com/frareb/aof). 255 We then averaged the values of breakpoint age obtained from two candidate time   We also compared the resulting breakpoint-based estimates of AOF with 281 values obtained from the more traditionally used, but arbitrary, threshold approaches.  (Table S1), the average time 296 required for emerged bees to perform their first foraging activity (Winston, 1994). We   Although the procedure was developed to detect the Age at Onset of Foraging (AOF) 334 in honey bees, this method could be of interest for other species. Thus, we carried 335 out a sensitivity analysis by simulating a broad range of virtual data to assess the 336 relevance and limit of our procedure. We first simulated two sets of data: (i) a "no 337 change simulated" scenario to assess the sensitivity of the procedure to detect

345
Assuming that the performance of the procedure to correctly detect (or not) 346 behavioural changes could be affected by the number of data points in time series 347 and the variance in data, we additionally modified these two parameters in both 348 scenarios. For that, we created a factorial design for which we sequentially increased 349 the number of data points (n) from 5 to 45 over 40 categories, and we sequentially    (Figure 2a). 376 GLMMs performed a posteriori (i.e. once the breakpoint was detected) 377 indicated that the AOF significantly differentiated two life-history stages, learning and 378 foraging, from their time-activity budget allocations (Figures 2c-e). Compared to the 379 learning stage, foragers performed more trips per day (GLMM with cohort as random  Table S2). The breakpoint-based method using Trip number also showed a 410 significant difference (t = 3.30, p = 0.003) as well as Trip time to a lesser extent (t = -411 2.33, p = 0.023; Table S2), suggesting that Trip number and Trip time should be 412 discarded as methods to estimate AOF in bees (Figure 3).  There was a non-linear relationship between AOF and the number of foraging 431 days (GLMM with cohort as random effect, F 3,862 = 24.771, p < 0.001; Figure 4a). 432 While the time lag between AOF and LSP, the so-called foraging performance, was 433 8.7 ± 4.1 days (Figure S2), AOF affected this duration quadratically rather than , that can bias the record of time-activity budgets. We recommend 508 using the simple hit-based AOF assessment method whenever the detection rate is 509 suspected to be low (e.g. < 95 %) and when such hit data are available (e.g. not

544
The aof automated method using individual life-long monitoring is generic 545 enough to be applicable to any tracking devices that record time-activity budgets in 546 free-ranging insects (i.e. harmonic radar, image-based optical counter, and RFID).

547
While developed for the study on honey bees, the present method could propose