ANGWSPERMS. 



417 



be seen even in the most rudimentary stage of development, and then a ' congenital ' 

 duplication is spoken of, commencing at the first inception of the organ, but this 

 expression only means that two rudiments make their appearance, where in other 

 flowers only one is found. But this fact may in many cases be partly referred to the 

 conditions of growth of the young organ l . It is a general rule, prevailing also in the 

 vegetative region of the plant, that the number of rudiments of organs increases on a 

 shoot for instance, if either the surface of the shoot continues of the same size but the 

 size of the rudimentary formations suddenly diminishes, or if the producing surface of 

 the shoot increases rapidly in size while the rudiments remain of the same size. An 

 instructive instance of the first case occurs in the perianth-leaves of the flower-heads of 

 Typha, which are arranged in two rows ; in the lower part of the head each young 



FIG. 346. Floral diagram of Capparideae. A Cleowe 

 droserifotia. B Polanisia gra-veolens. After Eichler. 



FIG. 347. Floral dia- 

 gram of the Cruciferae. 



perianth-leaf occupies half the circumference of the head at the point where it is inserted. 

 Towards the top of the head the perianth-leaves are smaller, and here instead of one 

 there are two or three rudimentary perianth-leaves quite separate from one another. 

 To speak in this case of a congenital splitting would be only a needless circumlocution. 

 The family of the Rosaceae supplies interesting conditions for the occurrence of similar 

 processes in its flowers. These are usually pentamerous and sometimes also tetrame- 

 rous. In some species the pentamerous corolla is succeeded by five stamens alternating 

 with the petals (Fig. 348 /). But in some species the rudiments of the stamens diminish 

 in size, and the first five stamens are followed not only by five but by ten other stamens, 

 which range themselves in the intervals (Fig. 348 //), and each pair of these ten young 

 stamens are closely connected with one of the five original stamens. I have shown at 

 some length in another place that it is impossible in this case to suppose a duplication 

 of five antisepalous stamens. In other Rosaceae (Fig. 348 ///) the same process takes 

 place on the first appearance of the stamens, that is, not five but ten rudiments of 

 stamens are formed alternating with the petals, and each pair of these are closely 

 connected with a rudiment of a petal. If the growth of the torus is everywhere uniform, 

 ten more stamens may make their appearance in front of the intervals between the first 

 ten stamens, and then we have an androecium consisting of two whorls- of ten members 

 each succeeding to a corolla of five members. But in many cases the growth of the 

 separate parts of. the torus after the formation of the rudiments of the first ten stamens 

 is not uniform. In Rubus Idaeus for instance the parts in front of the sepals grow more 

 vigorously than those in front of the petals, and consequently not one but several 

 rudiments of stamens are formed in front of each sepal quite unconnected with one 

 another ; it seldom happens that one only is formed, the torus having grown very little 

 in this part ; the number is usually two, and then one or two more are intercalated 

 (interposed) between them (Fig. 348 d\ but two remarkably small rudiments appear in 

 front of each petal, and these may be replaced by one larger one. In this case also 



1 Hofmeister, Allg. Morphol. d. Gewachse, p. 475. Schwendener, Mechanische Theorie d. 

 Blattstellungen, Leipzig, 1878. Goebel, Beitr. z. Morphol. u. Physiol. d. Blattes, No. Ill, Ueber die 

 Stellung d. Staubblatter in einigen Bluthen (Bot. Ztg. 1882). 



