Horsegram has been an important crop since the beginning of agriculture in many parts of South Asia. Despite horsegram's beneficial properties as a hardy, multi-functional crop, it is still regarded as a food of the poor, particularly in southern India. Mistakenly regarded as a minor crop, largely due to entrenched biases against this under-utilised crop, horsegram has received far less research than other pulses of higher status. The present study provides an updated analysis of evidence for horsegram's origins, based on archaeological evidence, historical linguistics, and herbarium collections of probable wild populations. Our survey of herbarium specimens provides an updated map of the probable range of the wild progenitor. A large database of modern reference material provides an updated baseline for distinguishing wild and domesticated seeds, while an extensive dataset of archaeological seed measurements provides evidence for regional trends towards larger seed size, indicating domestication. Separate trends towards domestication are identified for northwestern India around 4000 BP, and for the Indian Peninsula around 3500 BP, suggesting at least two separate domestications. This synthesis provides a new baseline for further germplasm sampling, especially of wild populations, and further archaeobotanical data collection.