Introduction

When I talked about how $F$-statistics could be partitioned into multiple levels, I imagined a situation in which we might have

• Inbreeding within local populations ($F_{IS}$).

• Differentiation among local populations within regions ($F_{SR}$).

• Differentiation among regions ($F_{RT}$).

I also pointed out that we can relate these statistics to the overall departure from Hardy-Weinberg through the relation

$$1-F_{IT} = (1-F_{IS})(1-F_{SR})(1-F_{RT}) \quad .$$