When agriculturists first began to cultivate wild legumes they observed that many seeds would not germinate within a comparatively short time after planting. Thus the problem of impermeable seeds began to demand attention many years ago. However, impermeable seeds are not confined to the Leguminosæ, as they occur also in the Malvaceæ, Chenopodiaceæ, Convolvulaceæ, Cannaceæ, and other families. Since the first account of the structure of legume seed coats by Malpighi (23 v. 1) in 1687, many investigators have contributed to our knowledge of the structure of the coats of seeds belonging to this family. Pammel (31) made an extensive study of legume seeds, including all the genera in the sixth edition of Gray's Manual, as well as genera not included in that publication. He found that the seed coat uniformly consisted of three layers, namely, the outer layer of Malpighian cells, the osteosclerid layer, and the inner layer of nutrient cells. Pammel's work included a study of the seed coats of _Melilotus alba_ and _M. officinalis_, and the descriptions and illustrations in his publication agree for the most part with the results obtained in the investigations reported in this article. However, more variation was noticed in the different layers of the seed coat than he describes. The cause of impermeability in seeds has been investigated by many. It has been found to be due to the embryos in some seeds, such as the hawthorns, but in most cases to the structure of the seed coat, and especially so in the Leguminosæ. Crocker (3) states that, exactly opposite to the common view, the cause of delayed germination generally lies in the seed coats rather than in the embryos. Nobbe (29) thought that the hardness of leguminous seeds was due to the Malpighian layer, and in a later publication Nobbe and Haenlein (30, p. 81) state that the absorbent power of many seeds is inhibited or entirely arrested by the cones of the Malpighian cells and the shields built up between them, which consist principally of cutin. Huss (15) agrees with Nobbe and Haenlein. Verschaffelt (39) found that the impermeability of the seeds of Cæsalpiniaceæ and Mimosaceæ investigated was due to, the inability of water to pass through the canals of the seed coat. By soaking the seeds in alcohol or other substances which change the capillarity of the pores, the seed coats were made readily permeable to water. Gola (6) states that the cause of the impermeability of seeds is the peculiar character of the Malpighian cells, which prevents their infiltration and consequent increase in volume, while Bergtheil and Day (2) found that the hardness of the seeds of _Indigofera arrecta_ was due to their possession of a very thin outer covering of a substance resistant to water. Ewart (5, p. 185) believes that in most impermeable seeds the cuticle prohibits the absorption of water, but gives as an exception _Adansonia digitata_, in which the whole integument seems to