One of the characters with which Mendel dealt in his hybridisation experiments with peas was, as is well known, the shape of the ripe seed. Weldon, in his criticism of Mendel’s interpretation of his results, showed that round were not discontinuously distinct from wrinkle peas, but that intermediate shapes connecting these two extremes were not infrequently exhibited. The answer which was made to Weldon’s criticism was that the intermediate shapes were due to spurious pitting or dimpling of the seed, and did not represent an intermediate condition of the germ which gave rise to them. And this answer was shown to be correct by the work of Gregory, who found that the starch-grains of round and wrinkled peas were quite distinct, and that they afforded an infallible test by which the real character of a pea with doubtful shape could be determined. Our knowledge of this subject has not advanced beyond the stage reached by Gregory in 1903; that is to say, we know no more about the inheritance of wrinkledness and roundness than Mendel did, except that each of these two characters is associated with a particular kind of starch-grain. What is the nature of the starch-grain in the hydrid; and how the characters of the starch-grains segregate, if they do so at all, in subsequent generations, are points on which we are at present ignorant. The observations, which I have to record, form the first instalment of an attempt to fill up this gap in our knowledge.