. 2019 Nov 14;37(3):923–932. doi: 10.1093/molbev/msz265

Table 1.

Inferred Island Fox Effective Population Sizes.

Population		${\hat{N}}_{e}$ (95% CI) Reported in Funk et al. (2016)		${\hat{N}}_{e}$ (95% CI) Using $σ_{D}^{2}$
San Miguel I.	13.7	$(13.2 - 14.1)$	15.3	$(14.5 - 16.1)$
Santa Rosa I.	13.6	$(13.5 - 13.7)$	13.3	$(13.0 - 13.6)$
Santa Cruz I.	25.1	$(24.6 - 25.5)$	22.8	$(22.4 - 23.3)$
Santa Catalina I.	47.0	$(46.7 - 47.4)$	40.9	$(40.4 - 41.6)$
San Clemente I.	89.7	$(77.1 - 107.0)$	59.1	$(53.0 - 66.7)$
San Nicolas I.	2.1	$(2.0 - 2.2)$	13.8	$(13.0 - 15.2)$

Note.—LD between unlinked loci provides an estimate for the effective number of (breeding) individuals in the previous several generations. Funk et al. (2016) used NeEstimator (Do et al. 2014) to estimate N_e for six island fox populations in the Channel Islands of California (left). We used this same data to compute N_e using our estimator for $σ_{D}^{2}$ instead (right), obtaining results largely consistent with Funk et al. (2016). Notably, Funk et al. inferred an extremely small size on San Nicolas Island ( ${\hat{N}}_{e} \approx 2$ ), whereas our estimate is somewhat larger and on the same order of magnitude of ${\hat{N}}_{e}$ from other islands with small effective population sizes. A 90% confidence intervals were computed via 200 resampled bootstrap replicates (see Materials and Methods).