“…For pollen data, only selected pollen types are shown and the pollen percentages are based on an upland pollen sum (lowland tree taxa of moist forests, light green; montane forest tree taxa [elevation range ∼2,000–3,000 m], dark green; grasses, yellow; high‐altitude shrub and herb taxa of the ericaceous and alpine vegetation belts currently largely restricted to elevations above 3,000 m [e.g., Hedberg, ], dark and light blue, respectively). From top to bottom: (1) Lake Albert pollen record (Beuning et al, ); (2) Lake Mahoma pollen record (Livingstone, ) and Mahoma glaciation of the Rwenzori Mountains based on dated moraines (Kelly et al, ); (3) lake surface temperature record of Lake Victoria (Berke et al, ) based on the TEX86 paleotemperature proxy; (4) Bwindi pollen record (MB3 core) (Marchant et al, ); (5) Muchoya pollen record (MC2 core) (Taylor, ); (6) Ahakagyezi pollen record (AH2 core) (Taylor, ); (7) Osokari soil erosion record represents radiocarbon dates of wood/tree trunks that were buried by hillwash in presumably open savanna woodland (Runge, ,b); (8) volcanic eruptions of Mount Karisimbi based on K‐Ar dates of lavas (De Mulder and Pasteels, ); (9) lake level record of Lake Kivu (Hecky and Degens, (dotted line); Zhang et al, (continuous line)), eastern gorilla subspecies divergence estimates (Roy et al, ), and scenario of eastern gorilla effective population sizes (i.e. reproducing individuals) based on Roy et al () for the past 10,000 years and possible earlier changes based on paleoenvironmental data; (10) Burundi highlands precipitation reconstruction based on pollen transfer functions from nine pollen sequences (Bonnefille and Chalié, ); (11) Lake Tanganyika deltaD leaf wax record (Tierney et al ) with more negative isotope values indicating wetter climatic conditions.…”