670 THE POPULAR SCIENCE MONTHLY. 



retrogressive development retrogressive, that is to say, if we regard 

 the lily family as an absolute standard: for the various alterations un- 

 dergone by the different flowers are themselves adaptive to their new 

 condition, though that condition is itself decidedly lower than the one 

 from which they started. The common rush and its immediate con- 

 geners resemble the lilies from which they spring in having several 

 seeds in each of the three cells which compose their pistil. But there 

 is an interesting group of small grass-like plants, known as wood-rushes, 

 which combine all the technical characteristics of the true rushes with 

 a general character extremely like that of the grasses. They have 

 long, thin, grass-like blades in the place of leaves ; and, what is still 

 more important, as indicating an approach to the essentially one-seeded 

 grass tribe, they have only three seeds in the flowei*, one to each cell 

 of the capsule. These seeds are comparatively large, and are richly 

 stored with food-stuffs for the supply of the young plantlet. One 

 such richly supplied embryo is worth many little unsupported grains, 

 since it stands a much better chance than they do of surviving in the 

 struggle for existence. The wood-rushes may thus be regarded as 

 some of the earliest plants among the great trinary class to adopt those 

 tactics of storing gluten, starch, and other food-stuffs along with the 

 embryo, which have given the cereals their acknowledged superiority 

 as producers of human food. They are closely connected with the 

 rushes, on the one hand, by sundry intermediate species which possess 

 thin leaves instead of cylindrical, pithy blades ; and they lead on to 

 the grasses, on the other, by reason of their very grass-like foliage, and 

 their reduced number of large, well-furnished, starchy seeds. 



In another particular, the rush family supplies us with a useful hint 

 in tracing out the pedigree of the grasses and cereals. Their flowers 

 are, for the most part, crowded together in large tufts or heads, each 

 containing a considerable number of minute separate blossoms. Even 

 among the true lilies we find some cases of such crowding in the hya- 

 cinths and the squills, or, still better, in the onion and garlic tribe. 

 But, with the wind-fertilized rushes, the grouping together of the 

 flowers has important advantages, because it enables the pollen more 

 easily to fix upon one or other of the sensitive surfaces, as the stalks 

 sway backward and forward before a gentle breeze. Among yet more 

 developed or degraded wind-fertilized plants, this crowding of the 

 blossoms becomes even more conspicuous. A common American rush- 

 like water-plant, known as eriocaulon, helps us to bridge over the gap 

 between the' rushes and such compound flowers as the sedges and 

 grasses. Eriocaulon and its allies have always one seed only in each 

 cell of the pistil ; and they have also generally a very delicate corolla 

 and calyx, of from four to six pieces, representing the original three 

 sepals and three petals of the lilies and rushes. But their minute 

 blossoms are closely crowded together in globular heads, the stamens 

 and pistils being here divided in separate flowers, though both kinds 



