SUPERPOSITION OF SUCCESSIVE PARTS. 199 



(whether adnate to or free from it) answer to one leaf which 

 has developed into two organs by a deduplication (372) taking 

 place transversely. This makes the inner and 

 antisepalons stamens the third floral circle 

 or the only truly andrcecial one, and sym- 

 metrically alternate with the petals on the one 

 hand and the carpels on the other. The 

 second hj'pothesis conceives that there is a 

 whorl suppressed between these antipetalous 

 stamens and the corolla : this, ideally restored, 

 gives symmetric succession and alternation to all the succeeding 

 whorls. The five glands in a Geranium-flower, alternate with and 

 next succeeding the petals (Fig. 384), were plausibly supposed to 

 represent this missing whorl, which according to Braun should be 

 an inner corolla ; according to others rather a primary circle of 

 stamens. The third is the recent hypothesis of Celakowsk}-, 

 which Eichler adopts : this regards the antipetalous stamens as 

 really the inner or second circle, and conceives that in the course 

 of development it has become external by displacement. The 

 difficulties of this Irypothesis are, first to account for this dis- 

 placement, and then for the anteposition of the carpels to the 

 assumed inner stamens in the great majority of these cases. 1 



1 In the first part of the Bliithendiagramme, Eichler inclined to the first 

 hypothesis, that of St. Hilaire (now very much abandoned on accoun^ of 

 the feeble evidence that there is any such thing as transverse or median 

 chorisis); in the second, he discards this in favor of Celakowsky's view 

 (published in Regensburg Flora, 1875). As to members which are morpho- 

 logically interior becoming exterior by outward displacement, Eichler cites 

 the staminodia or sterile stamen-clusters of Parnassia (Fig. 400, 401), and 

 the corresponding antipetalous stamens of Limnanthes, as clearly interior 

 in the early flower-bud, but exterior at a later period ; states that the vascu- 

 lar bundles which enter these stamens generally are either inner as respects 

 those of the episepalous stamens or in line with them ; that in some cases 

 (as in many Caryophyilaceae) the real insertion of the stamens is that of 

 direct diplostemony, while the upper part of their filaments and the anthers 

 are external to the episepalous series; that in most families with obdiplos- 

 temony examples of direct diplostemony occur, and still more cases with both 

 stamineal circles inserted in the same line ; and that, as a rule, the episep- 

 alous stamens are either later or not earlier formed th:in the epipetalous. 

 As to the position of the carpels before antipetalous stamens and petals, 

 Celakowsky suggests that this may result from the outward recession of 

 those stamens affording more room there, while in the normal case the 

 greater space is over the episepalous stamens. And, indeed, exceptions 

 to the prevalent position are not uncommon both in direct diplostemony 



FIG. 384. Diagram (cross- sect ion) of the flower of Geranium macnlatum, exhibiting 

 the relative position of parts, and the symmetrical alternation of circles, i.e. sepals, 

 petals, greenish bodies called glands, antipetalous stamens, antisepalous stamens, 

 carpels. 



