Table 2. Estimates of (a) social transmission effects for LS re-use (RU1) and moss-sponging (M) variants, giving the multiplicative effect on learning rate of each observation (1×, no effect); (b) the ratio of social transmission effects between M and RU1; and (c) the estimated number of acquisitions that were by social transmission, excluding the innovation event.

	(a) Social Transmission (Multiplicative Effect Per Observation)	(b) Ratio: M Effect/RU1 Effect	(c) % of Events by Social Transmission
RU1	1.07×(0.58–2.48)	—	3% (0%*–19%)
Moss-sponging KW included	14.93×(4.67–88.24)	2.42×(4.67–72.24)	85% (80%–86%)
Moss-sponging KW excluded	21.17×(4.19–679)	15.90×(3.00–230)	99% (92%–100%)

Estimates are model-averaged estimates, with unconditional confidence intervals in parentheses. For M, estimates are given both with KW included (conservative estimate) and excluded (see text for explanation).

*Note that the lower 95% C.I. limit for the social effect on RU1 is <1, meaning each observation decreases the rate of learning; we set this situation to be zero events by social transmission.