DICOTYLEDONS, c5^ 



arranged in whorls, or the flowers are cyclic ; only in a comparatively small number 

 of families (Ranunculacese, Magnoliaceae, Calycanthaceae, Nymph^acese, and Nelum^ 

 biaceae) arc all or some of them arranged spirally {acyclic or hemicyclic^). 



In Cyclic Floiuers the whorls are usually pentamerous, less often tetramerous, 

 both numbers occurring in nearly-related plants. Dimerous or trimerous, or combi- 

 nations of dimerous and tetramerous whorls are much less common than penta- 

 merous, and are usually characteristic of smaller groups in the natural system. 



When the floral whorls are tetramerous or pentamerous, they are generally four 

 in number, and are developed as Calyx, Corolla, Androecium and Gynceceum. In 

 dimerous or trimerous flowers the number of the whorls is much more variable, and 

 then it is not uncommon for each organ to be made up of two or three whorls ; 

 while in the previous case the multiplication of the whorls is almost entirely confined 

 to the androecium. 



The corolla is frequently absent, and the flow^ers are then said to be apelaloiis. 

 When the calyx and corolla are both present the number of their parts (sepals and 

 petals) is almost always the same (Papaver is an exception) ; but this is not the case 

 with the number of the whorls. In Cruciferae, for example, the calyx consists of two 

 decussate whorls of two sepals each, the corolla of one whorl of four petals. When 

 the perianth and androecium are both present (whether the former consist of calyx 

 only or of both calyx and corolla), the number of their parts is usually the same, or 

 the flower is isosle7?iofwus, but the stamens are often more, rarely fewer in number 

 than the parts of the perianth, and the flower is then anisosiemonoiis. When the 

 flower is tetramerous or pentamerous the number of carpels is usually less ; when the 

 flower is dimerous or trimerous, or when the parts are arranged spirally, the number 

 of carpels is not unfrequently larger. 



It will be seen from this brief outline that the relations of number and position 

 in the parts of the flowers of Dicotyledons are very various, and cannot be referred, 

 as is the case with ^lonocotyledons with but few exceptions, to a single type. Even 

 the establishment of difl'erent types for the larger groups is attended with great 

 uncertainty, since the knowledge of development necessary in order to refer par- 

 ticular forms of flowers to general formulae is often wanting. The too universal 

 application of the spiral theory to phyllotaxis even in the case of cyclic flowers 

 has often increased the difficulty, and has even occasioned doubts which would 

 not have arisen without the theory. 



For the great majority of Dicotyledons the floral formula may be given 

 6'»/'„6'/,,(^„^...) C„(_„i). This formula holds good for most pentamerous flowers 

 and for those which are truly tetramerous (or octamerous as Michauxia) ; so that n 

 is in these cases 5 or 4 (or 8 as the case may be). In the androecium an indefinite 

 number of (alternating) whorls ^/,(+„4....) must be assumed in order to include 

 the large number of flowers in which the androecium consists of more than one 

 whorl (as^.^. Fig. 420). The mode expressing the gynaeceum C„(_^) is intended 

 to show that very commonly the number of carpels is fewer than 5 or 4 (or 8 as the 

 case may be) ; m may be of any value from o to «. In the majority of gamopetalous 



Compare pp. 523 and 531. 



