SEXUAL REPRODVCTION. 



301 



seeds, it ver)- often does not open at maturity. Consenuently, 

 the seed-coats ordinarily remain thin, and the protective 

 function is put upon the pericarp. In some cases the carpel 

 becomes adherent at an early stage to the surface of the 

 ovule, and at maturity the pericarp is so firmly attached that 

 it can scarcely be distinguished from the seed-coats them- 

 selves. Such a change takes place in the fruit of most grasses, 

 and the grain so formed is ordinarily mistaken for a seed 

 (fig. 346). When dry fruits are one-seeded and indehiscent 



12 34 



Fig. 346. — A bmall p rliun fr m Che margin of a transverse section of grain of oats, 

 T, 2, pericarp 3 seed coats 4, remains of tile sporangium ; 5-7, endosperm ; 5, 

 gluten cells , 6, cells containing large compound starch-grains (compare tig. 174) at 

 7 richer in gluten, with less starch. Magnified about 325 diam. — After Harz. 



the pericarp usually bears whatever special contrivances are 

 necessary for the distribution of the seeds. (See further ^ 

 4S9 ff. ) If, howe\er, the pericarp contains many seeds, it 

 generally breaks at maturity to allow the loosened seeds to 

 escape. The extent and position of the opening into the 

 seed chamber or cliambers are extremely various. In some 

 cases the openings are so small as to be mere slits or pores 

 (fig. 347). In others a more or less circular line of breakage 

 forms a little door or valve which opens and closes with 



