Abstract Leaf area index (LAI) is an important indicator of canopy development and function, but is time consuming to measure directly. In this paper the authors investigate an indirect and non-destructive method of determining LAI in field-grown Pima cotton ( Gossypium barbadense L.) planted in rows, using a commercially available instrument (LiCor LAI-2000, Plant Canopy Analyzer (PCA); Lincoln, NE, USA). The PCA uses a circular sensor head, with a full 360° field of view, consisting of five sensor rings, each viewing a narrow range of zenith angles. Excellent agreement was obtained between destructively determined LAI and indirectly measured LAI, over a range of LAI induced by irrigation treatments. However, the near 1:1 relationship between the two methods required specific measurement and calculation protocols. All measurements were obtained at dusk. Only a 90° sector (25%) of all five sensor rings was exposed, directed along the rows and, when possible, away from the setting sun. Measurements were obtained at four positions across the row/inter-row, which were averaged to characterize the canopy. Underestimation of LAI was observed when data from all five sensor rings were used in the calculation. This bias was removed when data from only the central three, most vertical, rings were used. This protocol empirically compensated for significant violations of the theory underlying the indirect measurement technique. As most row crops similarly violate the theory, this protocol could have widespread utility.