862 



SCIENCE 



[N. S. Vol. XLII. No. 1094 



ledonous or dicotyledonous. My prophecy 

 is that they will be found to be both. A 

 study of cotyledony in general has shown 

 that these two conditions, and also mono- 

 eotyledony, are merely different, and often 

 variable iinal expressions of a common 

 method of development. A cotyledonary 

 zone or ring always develops around the 

 growing point, and upon this ring a vari- 

 able number of primordia appear, nearly 

 always more than finally develop. The 

 whole ring continues to develop in connec- 

 tion with one or two or more growing 

 points, the others having been cheeked by 

 conditions easily explained by the ontogeny 

 of the embryo, especially by the time and 

 position of the appearance of the immedi- 

 ately succeeding members. When one finds 

 not only dieotyledony, but also polycoty- 

 ledony, among the Monocotyledons, it be- 

 comes apparent that the number of coty- 

 ledons is a variable. The wonder is that it 

 is as constant as it appears to be. 



The interesting evolutionary feature is 

 that polycotyledony is so much more com- 

 mon among gymnosperms than among 

 angiosperms. It is perhaps safe to say that 

 it was as common among most primitive 

 gymnosperms as was dieotyledony; or 

 rather that the number of cotyledons was 

 much more variable than in any living 

 group of seed plants. This interesting situ- 

 ation is still further emphasized by the re- 

 markable constancy of dieotyledony in the 

 Dicotyledons and of monocotyledony in the 

 Monocotyledons, but I know of none of them 

 in which less than four cotyledonary grow- 

 ing points start. The conditions that seem 

 to determine the number of cotyledons to be 

 developed by a cotyledonary ring are too 

 numerous to be discussed here, but in gen- 

 eral they have to do with the rate of growth 

 of the subsequent members of the embryo. 

 For example, if the subsequent leaves begin 

 to appear almost immediately and develop 



vigorously, the cotyledonary ring usually 

 becomes one-sided in development, and the 

 result is a single large cotyledon in an ap- 

 parently terminal position. Many mono- 

 cotyledonous embryos, in which for some 

 reason there is an elongation of the stem 

 before the first leaves begin a vigorous 

 growth, develop two cotyledons, as in the 

 case of numerous grasses. This is the usual 

 sequence in Dicotyledons ; while in polycot- 

 yledonous forms there is much delay in the 

 appearance of the subsequent members, and 

 no inhibition of cotyledon primordia. All 

 this variation in the number of cotyledons 

 suggests variations in the conditions of 

 growth, since it depends upon rate of 

 growth in the so-called "plumule." 



4. The Seed. — Another noteworthy illus- 

 tration of progressive evolution in gymno- 

 sperms, associated with the same conditions 

 that seem to have determined the changes 

 previously cited, is the progressive simpli- 

 fication of the ovule and seed. As yet the 

 most primitive ovule is not available, and 

 the hiatus in our knowledge between the 

 fern sporangium or sorus, and the most 

 primitive known ovule is complete. In 

 that unknown region heterospory devel- 

 oped, and then the seed-forming ovule, but 

 the steps are left to conjecture. The 'in- 

 teresting fact, however, is that the most 

 primitive o\Tales and seeds we know are the 

 most complex, and that there has been a 

 progressive simplification through the 

 whole series of seed plants. This simplifi- 

 cation has not only involved the layers of 

 the testa, as often pointed out ; but its grad- 

 ual progress is most completely shown by 

 the vascular supply to the ovule and seed. 

 The very gradual elimination of the vas- 

 cular elements is a measure of the pro- 

 gressive simplification of the whole struc- 

 ture. We have found that the vascular 

 supply does not determine the structure; 

 but the structure determines the vascular 



