Remotely sensed imagery is an important tool for monitoring changes in land cover and is becoming increasingly important for quantifying terrestrial plant biomass. Multi-temporal studies tracking changes in woody plant cover in arid rangelands within the United States are limited to early aerial photography for base-line conditions. For the Santa Rita Experimental Range (SRER) in southeastern Arizona, panchromatic aerial photography from 1936, commissioned by the U.S. Soil Conservation Service, is the source of earliest synoptic aerial depictions of vegetation cover in the form of black and white photographs at a scale of 1:31,640. We capitalized on the well-documented land use history for the SRER to generate a spatially-explicit database of velvet mesquite (Prosopis velutina) stand structure within two 1.8-ha field plots established by W. McGinnies in 1932 to answer two questions: (1) What are detection limits of 1936 aerial photography (i.e., smallest reliably discernable P. velutina canopy)?, and (2) How much P. velutina biomass do we miss using 1930s aerial photography? An allometric relationship between canopy area and aboveground woody biomass was applied to 1932 P. velutina canopy measurements to translate detection limits of 1936 aerial photography to biomass. Shrub polygons were manually digitized on geo-coded 1936 imagery and linked to co-registered McGinnies’ 1932 field measurements of P. velutina canopy area. Based on a conservative shrub size detection threshold (90^th percentile for plants not visible on imagery) of 3.8-m² ca. 5.8% of P. velutina biomass was missed. When additional sources of error (e.g., spatial co-registration and observer bias with aerial photo interpretation) were taken into account, the percentage of biomass missed increased to 17.5%. These analyses illustrate the importance of standardizing aerial photo interpretation protocols and accounting for uncertainties associated with woody plant biomass estimates derived from remotely sensed imagery.