Table 2. Results of a pair-based analysis of when ant pairs initiate trophallaxis.

We modeled trophallaxis initiations for each pair as coming from an inhomogeneous Poisson point process, where the rate of trophallaxis initiations depends on colony and local-density effects. We included effects for each colony and experimental condition (high or low), and also considered interactions between colony and the local density at 5 mm (n5), 10 mm (n10), 15 mm (n15), and 20 mm (n20).

	Estimate	Standard error	Z-value	p-value
Colony 1	−5.80	0.07	−82.25	<2e-16
Colony 2	−6.04	0.09	−69.18	<2e-16
Colony 3	−8.59	0.29	−30.04	<2e-16
Treatment	−0.49	0.04	−11.11	<2e-16
n5	0.18	0.04	4.08	4.60e-05
n10	0.05	0.02	2.08	0.04
n15	−0.41	0.03	−15.01	<2e-16
n20	−0.26	0.02	−10.57	<2e-16
col2:n5	−0.04	0.07	−0.62	0.5335
col3:n5	0.40	0.07	5.98	2.30e-09
col2:n10	−0.03	0.04	−0.69	0.49
col2:n15	0.08	0.04	1.77	0.08
col3:n15	0.19	0.03	5.84	5.29e-09
col2:n20	−0.002	0.04	−0.06	0.96
col3:n20	0.07	0.03	2.14	0.03