yan. I, 1874] 



NA TURE 



165 



a net, the anthers remain filled with pollen-grains, and 

 never, or only exceptionally, come spontaneously into con- 

 tact with the stigmatic branches, the free ends of their 

 filaments at a later period bending downwards, and the 

 branches of the styles, remaining considerably above them 

 (Fig. 24). Conversely in the flowers of M. rotundifolia, 

 when the visits of insects are prevented, the anthers, 

 filled with pollen-grains, remain in so high a position, 

 and the stigmatic branches bend so far downwards as to 

 come abundantly into contact with the pollen-grains, self- 

 fertilisation being thus inevitable (Fig. 26). 



Epilohiidii angiisiifoliuin and parvifloruin 



differ most strikingly in a similar manner. The flower 

 of E. aii_i^!/s/ifoliuiii, being of larger size, brighter co- 

 lour, grouped in loner splendid clusters, and exciting- 

 attention at a great distance, are so largely visited and 

 cross-fertilised by insects'* as never to have need of self-fer- 

 tilisation, which has actually become impossible ; the four 

 stigmatic branches unfolding so long after the maturity 

 of the eight anthers, and so far overtopping them, as 

 to be completely shut out from the pollen of the same 

 flower. The flowers of E. parvifloruin, on the other 

 hand, being of smaller size, lighter colour, and single, 

 are so inconspicuous that insects but very rarely visit them. 

 Accordingly, its four upper anthers so closely surround 

 the four-lobed stigina, which is mature at the same time, 

 as to cover it largely with their pollen, whilst the pollen- 

 grains of the four lower anthers lying on the way to the 

 honey, cannot reach the stigma of the same or of another 

 flower unless transferred by insects. 



Polygonum 



Among the many species of the genus Polygonum 

 which grow in our country there are two, /'. Fagopyruiii and 

 Bistorta, most distinguished by their attractiveness for 

 insects, which is due not only to the size and colour of 

 the single flowers and to their collection into handsome 

 spikes, but also, and even more perhaps, to their abun- 

 dance of honey secreted by eight globular nectaries at 

 the base of the filaments («, Figs. 26, 27). With re- 

 ference also to the frequent visits paid them by insects, t 

 these two species have been adapted to inevitable cross-fer- 

 tilisation by their visitors, self-fertilisation having at the 

 same time become difficult or almost impossible. The 

 manner in which this advantage has been attained being 

 very different in the two species, it is evident that in this 

 case the adaptation to cross-fertilisation by the visits of 

 insects cannot have been inherited from the common 

 parents of the genus, but must have been acquired by the 

 single species during their evolution. 



P. Fagopyruni has acquired, as shown in Figs. 26 and 

 27, the same kind of dimorphism which has been so fully 

 explained by Darwin in Primula J and Linum.§ In 

 both of the two kinds of flowers (which occur only on 

 difl'erent plants) there are three styles and eight stamens, 

 three of the stamens closely surrounding the styles and 

 opening outwards, the five others inserted more outwards, 

 alternating with the leaves of the perianth and opening 

 inwards. An insect, therefore, visiting a flower for honey 

 and pushing its head or proboscis between the inner and 

 outer stamens into the base of the flower, cannot avoid 

 being charged with pollen, especially in those parts of 



* On the flowers of EJfilohium aHg;ustifoliutn I have hitherto observed 26 

 species ol insects, T4 of them belonging to the family of bees, many of them 

 very frequently : on those oi E. parz'ijloritm I found only once Meligethes, 

 and once a butterfly {Picris raptr- L.) repeatedly sucking the honey of its 

 flowers. 



t On the flowers of P. Fagopyntm I have observed 41 species of insects, 

 among them 2r Diptera and 12 ApidK ; on the flowers of P. Bisiorta 18 

 species of insects, among them 9 Diptera and 3 ApidEe ; many of the visitors 

 of each species very frequently. 



J On the two forms or dimorphic condition in the species of Primula and 

 their remarkable sexual relations (Proc. of the Linn. Soc. vi. (1862) ; Bot. 

 PP- 77-79- . 



§ On the existence of two forms and their reciprocal sexual relation in 

 several species of the genus Linum, Ibid. 1863, pp. 69-83. 



its body which, whilst it is sucking the honey, are pressed 

 against the anthers. Now, the place occupied in one of 

 the two kinds of flowers by the anthers, is occupied in the 

 other kind by the stigmas, the same parts of the body of 

 the insect which in the long-styled form were pressed 

 against the anthers, come into contact in the short-styled 

 with the stigmas, and conversely. Thus it is inevitable 

 that insects effect chiefly what is called legidmate fer- 

 tilisation, i.i\ transmission of the pollen of the long-styled 

 flowers to the stigmas of the short-styled, and of the pollen 

 of the short-styled to the stigmas of the long-styled form. 

 Fertilisation by pollen of the same form, however, and 

 even of the same flower, is not impossible, and in the 

 short-styled fl wers even spontaneous self-fertilisation 

 may happen, by pollen-grains falling down from the anthers 

 upon the stigmas. 



The same advantage which P. Fagopyruin has attained 

 by dimorphism (Darwin) or heterosfyly (Hildebrand), has 

 been gained in the flowers of P. Bistorta b> protan- 

 drous dichogamy, i.e. by the anthers so far preceding in 

 their development the stigmas that in the first period of 

 the flower (Fig. 2S) only mature anthers, at a later period 

 (Fig. 29) only mature stigmas are present, the anthers 

 having then commonly fallen off. It is readily seen that 

 such flowers also, when perseveringly visited by insects, 

 are always inevitably intercrossed, no other mode of the 

 transmission of poden being possible than from younger 

 flowers to the stigmas of older ones. It is only when the 

 visits of insects are completely wanting during the first 

 period and the anthers remain clothed with pollen while 

 the stigmas attain their maturity, that self-fertilisation by 

 insects or even spontaneous self-fertilisation is possible. 



The least attractiveness for insects, on the contrary, 

 among all native species of Polygonum is possessed by 

 P. aviculari:, its flowers (Figs. 30 and 31) being of small 

 size, of greenish and white or reddish colour, standing 

 singly on procumbent plants and offering only a small 

 quantity of pollen to insects, but, as far as I have been 

 able to see, no honey. No wonder that insects are in- 

 duced only in very rare cases to visit and fertilise them,* 

 and that, in compensation for the loss of cross-fertili- 

 sation, these little flowers regularly experience spon- 

 taneous self-fertilisation, the three inner anthers lying so 

 close to the stigmas that their pollen-grains inevitably 

 come into contact with them (Figs. 30 and 31). 



Of the many other native species of Polygonum, which 

 are all intermediate, as to their attractiveness for insects, 

 between those now described, I will only remark briefly 

 upon P. Pcrsico.ria, which is of more especial interest 

 because of its flowers presenting great differences of 

 structure. In this species, instead of eight nectaries there 

 are only five developed, and these secrete a much smaller 

 quantity of honey than those of /'. Fagopyruin 

 a.nd Bistorta. Its spikes of flower, moreover, being less 

 conspicuous than in those species, the visits of insects 

 are somewhat rare, even in sunny weather, although far 

 more frequent than in P. avifutarc-.f Fertilisation by 

 insects, consequently, is by no means secured. Cor- 

 responding to this uncertain agency of insects the sexual 

 organs of the flower are in a remarkably fluctuating con- 

 dition, undecided, as it were, between adaptation to cross- 

 fertilisation by the visits of insects, and to self-fertilisation. 

 Thus, of the eight stamens, sometimes only the five outer 

 ones are developed, the three otheis being reduced to rudi- 

 mentary filaments ; and this condition is apparently the 

 most favourable to cross-fertilisation, as any honey-seek- 

 ing insect must touch the anthers in every flower with 

 one side of its proboscis, the stigma with the opposite 

 side, to which it thus cannot fail to transfer pollen-grains 



* After having repeatedly in vain watched P. ainculare in very hot sunny 

 noons of the month of August 1S71, I succeeded in observing some small 

 'iyif\\\ix^Ascia fodagrica F., Syritta pipiens L., and Metithreptus men- 

 //rrt.j/W L.) visiting its flowers. 



t I have observed in the flowers of P. Persicaria altogether 11 species of 

 insects, among them 7 Diptera, and these as the most frequent visitors. 



