GoKawiila, Experimental design. A tracking study using small cohorts of bees from three observation colonies was conducted. A total of 18,000 newly emerged bees from the same apiary were marked with different colours (white, blue, yellow, red, green, and pink) to indicate their age. Specifically, 3,000 bees were marked with a single colour each day, and 1,000 of these bees were randomly introduced into each of the three observation colonies. Following a 15-day preparation stage, observations began on day 20, at which point the marked bees corresponded (white: 20 d, blue: 17 d, yellow: 14 d, red: 16 d, green: 13 d, and pink: 5 d). To induce WC construction, a newly built wax foundation was added to the colonies. Simultaneously, royal jelly production strips with half-built queen cell cups were inserted to induce QC construction. The observation period lasted for 10 days. On day 25, bees exhibiting QC and WC construction behaviours were captured from the hives for further analysis. b, Bee behaviours were analysed from recorded videos. Workers engaged in QC or WC construction were identified by colour to estimate their ages, allowing the proportion of each age cohort participating in these tasks to be calculated relative to the total number of marked bees. A two-sided chi-squared test was applied to compare the colour distributions between the two behaviours (χ2 = 29.23, df = 5, P < 0.001). c, Task allocation of worker bees of different ages to QC and WC construction. The left y-axis indicates the relative percentage of bees performing specific tasks, while the right y-axis displays the trend values fitted using LOESS smoothing. Two-sided linear mixed models (LMMs; n = 3 biologically independent colonies) were used to evaluate the effects of age, task, and hive on these relative percentages. The model was defined as: Value ∼ Age × Task + (1|hive), incorporating age, task, and their interaction (Age × Task) as fixed effects, and the beehive as a random effect. Exact F and P values for the fixed effects are shown in the panel. The gap in data on day 28 was due to adverse weather conditions preventing observation. d,e, Relative expression of key wax synthesis genes: LOC724560 (fatty acyl-CoA reductase 1 like gene, FAR), and LOC113218558 (acyl-CoA Δ11 desaturase–like gene, SCD). Boxplots represent relative gene expression values, with lines linking data from the same experimental colony (Colony 1: circles; Colony 2: squares; Colony 3: diamonds) across the two construction behaviours (QC: blue boxes; WC: orange boxes). Box plots show the median (centre line) and interquartile range (IQR, box limits). Whiskers extend to 1.5 × IQR, and individual data points are shown with jitter. Different letters (e.g., a, b, c) indicate statistically significant differences (adjusted P < 0.05). n = 3 biologically independent colonies for each age and task combination. The ridgeline plots (right panels) display the data distribution for bees performing each task (n = 18 biologically independent samples per task), where the x-axis indicates gene expression levels and the y-axis reflects the density (relative proportion) of samples at each expression level.
Source data