The precipitation recycling ratio is the fraction of rainfall over a given area of the land surface that entered the atmosphere as evaporation from the same area. Recycling is a feedback in the most concrete sense of the word, in that a physical mass (water) is passed back and forth between land and atmosphere. Thus, we might expect recycling to have a connection to land surface memory – the propensity for anomalies in soil wetness or vegetation to persist and potentially reinforce local and regional weather and climate anomalies. Land surface memory is quantified here by the decorrelation time scale for soil wetness. We investigate the character of this connection through comparison of soil moisture memory calculated from the global data sets of the Second Global Soil Wetness project (GSWP-2) with estimates of precipitation recycling ratio computed using a quasi-isentropic back-trajectory analysis based on observed precipitation and 25 years of atmospheric reanalyses. Geographic and seasonal variations in the connection between precipitation, recycling and memory show strong relationships across many parts of the world, with the nature of the relationships differentiated by hydrologic regime. In monsoon regions there are strong inverse relationships between rainfall and recycling, as the source of wet season rains is advected from elsewhere, while most of the major agricultural zones of the world have strong positive relationships. Recycling ratios are very static and confined to a narrow range of values in rainy areas, but can vary greatly in space or time in drier regions. Long memory is largely restricted to these arid and semi-arid regions.