Oct. II, 1877] 



NATURE 



507 



FER TILISA TION OF FLO WERS B Y INSECTS ' 



XVII. 



Abortion of ail tlie Stamens in a Flower in Four Successive 



Periods 



T N the theory of the development of the organic world 

 •*■ useless and aborted organs are always of especial 

 interest, as no other plausible explanation of them can be 

 given except that they are inherited from ancestors to 



which, in other conditions of life, they were useful ; it 

 may therefore be worth referring to a flower in which, in 

 four successive periods, all stamens have been aborted, 

 and accordingly four different degrees of abortion are to 

 be distinguished. The species in which these flowers are 

 found. Salvia pratensis, is a very common one, but the 

 flowers alluded to either do not occur at all in the usual 

 habitats of this species, or have hitherto been over- 

 looked by most botanists. I found them during my last 

 excursions in the Alps in some valleys of Switzerland 



Ficf.ua 



^< 



£iij:U1 



. 116-129. — Salvia pratensis.^ Fig. 116 —Side view of a hermaphrodite fl 

 a female flower (si : i). Fig. 118 —Lower part of the two stara 

 side (7 : i). Fig. 119. — Front view of the same. FlG. 120 - 

 hidden behind the connective. Fir.. I2t.— Right stamen as si 

 the calyx and of the corolla having been removed (7 : i). Figs. 12.S-129 



(Albula, Julia, Landwasser, and Landquart valley). In 

 these, and probably many other valleys of the Alps up to 

 1,200-1,400 metres above the sea-level, besides the usual 

 stems of Salvia pratensis with large hermaphrodite 

 flowers, other stems with smaller purely female flowers 

 are by no means rare. In these localities, consequently. 



Fii)l20 



Salvia pratensis is in the same state as Glechoma,' 

 Thymus, and some other Labiatae in all or most of their 



» Continued from vol. xv. p. 475. 



' In all fieu'es ^rt = calyx, cr(7 = corolla, «== nectary, «» = ovary, j = 

 ftyle, i/ = stigma, jr = ruoiments of the two aborted upper stamens, _/f = 

 filaments of the two lower stamens, ct = connective, « = upper anther-cell, 

 / = lower anther-cell. / = point of union of the two metamorphosed lower 

 anther-cells, pi — point of the filament on which the connective rotates. 



3 See Natukb, vol. viii., pp. 121, 143, 161. 



r with the corolla partly removed (35 : i). Fic. 117 —Side view of 



maphrodite flower viewed obliquely from the front and from the right 



part of the le't stamen alone .seen on the inner side ; the filament being 



[he outside (7 : i). Fig. 12a.— Side view ol a female flower, the half of 



Gradations of abortion of the two last stamens (7:1). 



habitats ; Salvia- pratensis is here, as Mr. Darwin 

 calls it, in his late work,' a gynodia;cious plant. In all 

 other gynodicecious LabiatEe two abortions of stamens 

 have occurred in two successive, periods ; in Salvia 

 pratensis, as I shall show, four. 



I. The LabiatEe, as well as the Scrophulariaceje, have 

 apparently descended from plants with five stamens. 

 But as soon as the common ancestors of the Labiate 

 family adapted their flowers to cross-fertilisation by 

 bees in such a manner that their stigmas and anthers 

 must necessarily be touched by the backs of these visitors, 

 the uppermost of the five stamens stood in the way of 

 the style, which for the purpose of this cross-fertilisation 

 must stretch along the middle line of the upper side of 

 the corolla and bend one of its two stigmatic branches 

 downwards. Thus the uppermost stamen having become 

 not only useless, but even directly disadvantageous, was 

 doomed to abortion, and in the long time that has elapsed 

 since then, has been so completely eliminated by natural 

 selection, that not the smallest trace of it has reinained, 

 and only very exceptionally does it reappear by atavism. 

 In those Labiatje in which the adaptation described has 

 been perfected, the reappearance of the fifch stamen hap- 

 pens, indeed, so extremely rarely that I have only once had 

 the opportunity of seeing it, in a single flower of Lamium 

 album, in which the upper lip was wanting, and, instead 

 of it the fifth stamen was present. In the flowers of 

 Mentha, however, in which the peculiarities of the Labiatae 

 just-mentioned are much less developed, the fifth stamen, 

 as I am informed by Dr. E. Krause, of Berlin, reappears 

 more frequently. 



' Oa the Dififerent Forms of Flowers in Plants of the Sa 



: Specie 



