In 1980, Wildman et al. (Bot Gaz 141: 24-36) proposed a three-dimensional model for chloroplast structure whereby the grana were arranged in non-overlapping rows, like beads on a string. This string-of-grana model was developed from phase microscope analysis of living cells and partially disrupted, isolated chloroplasts. However, models based on analyses by various electron microscope (EM) techniques (which inevitably encompass a relatively small fraction of the whole chloroplast) indicated that grana are interconnected in all directions by intergranal lamellae and not just along a single 'string.' Hence the string-of-grana model was not widely accepted. Recently, confocal laser scanning microscopy (CLSM) of both living and fixed cells, which gives views of the three-dimensional disposition of grana by imaging Photosystem II fluorescence over much larger sample volumes, that is, the entire chloroplast, has revealed that, although many grana are apparently not in any discernible arrangement, some are indeed present in strings of varying lengths in a range of taxa. The topic therefore warrants revisiting, using techniques, for example, such as EM tomography to assess the degree of variation in the geometry of intergranal connections in whole chloroplasts, and its possible functional consequences and developmental origins.