Climate records for locations across the southern slope of the Himalaya between 77°E and 91°E were selected together with discharge measurements from gauging stations on rivers draining partially-glacierised basins tributary to the Ganges, with a view to assessing impacts of climatic fluctuations on year-to-year variations of runoff during a sustained period of glacier decline. The aims were to describe temporal patterns of variation of glaciologically- and hydrologically-relevant climatic variables and of river flows from basins with differing percentages of ice-cover. Monthly precipitation and air temperature records, starting in the mid-nineteenth century at high elevation sites and minimising data gaps, were selected from stations in the Global Historical Climatology Network and CRUTEM3. Discharge data availability was limited to post 1960 for stations in Nepal and at Khab in the adjacent Sutlej basin. Strengths of climate–runoff relationships were assessed by correlation between overlapping portions of annual data records. Summer monsoon precipitation dominates runoff across the central Himalaya. Flow in tributaries of the Ganges in Nepal fluctuated from year to year but the general background level of flow was usually maintained from the 1960s to 2000s. Flow in the Sutlej, however, declined by 32% between the 1970s and 1990s, reflecting substantially reduced summer precipitation. Over the north-west Ganges–upper Sutlej area, monsoon precipitation declined by 30–40% from the 1960s to 2000s. Mean May–September air temperatures along the southern slope of the central Himalayas dipped from the 1960s, after a long period of slow warming or sustained temperatures, before rising rapidly from the mid-1970s so that in the 2000s summer air temperatures reached those achieved in earlier warmer periods. There are few measurements of runoff from highly-glacierised Himalayan headwater basins; runoff from one of which, Langtang Khola, was less than that of the monsoon-dominated Narayani river, in which basin Langtang is nested.