396 AUToi.A.MY. 



Amongst the contrivances mentioned in the last chapter as being adapted to 

 brmg about cross-fertilization at the commencement of a flower's period of bloom, 

 was the production of heterostyled flowers (see pp. 302 and 312). It was pointed 

 out (p. 316) that in heterostyled plants the very first, or last (as the case may be), 

 of all the flowers of a particular species are, in consequence of the dichogamy which 

 prevails, fated to be crossed with flowers of another species, or, in other words, to 

 undergo hybridization, and that, according to experience, this crossing is often 

 effective. Considering the results obtained in the cases of other plants with herma- 

 phrodite flowers, one would expect to meet with some provision for the timely fulfil- 

 ment of autogamy in heterostyled species as well, i.e. that in the event of no insects 

 visiting a plant of the kind, the stigmas should at the proper moment lie dusted 

 with pollen from the anthers in the same flower. This expectation has been, in 

 fact, confirmed; all the investigations directed to this question having resulted in 

 showing that a process of autogamy takes place also in heterostyled flowers, but is 

 always confined to one only of the forms which together constitute the species. In 

 one section of the heterostyled species, including, for example, those which belong 

 to the genera Gentiana, Menyanthes, and Thesium, the short-styled flowers are 

 adapted to autogamy, whilst in others, such as the heterostyled species of .1/. rt< asia 

 and Pulmonaria, autogamy takes place in the long-styled flowers. In Primula 

 longiflora and P. minima it is the short-styled flowers which have their stigmas 

 dusted with pollen from their own anthers, whilst in Primula Awricula, and /'. 

 glutinosa it is the long-styled flowers which thus accomplish self-fertilization. All 

 these variations are exactly adjusted with reference to the other arrangement- for 

 promoting autogamy in the flowers in which they are exhibited. 



( )ne contrivance which deserves mention here is the disparity in the size and 

 conspicuousness of the floral envelopes in the two forms of flowers of the same 

 species. In Primula longiflora and P. minima the long-styli 1 flowers, and in 

 Primula Auricula and P. glutinosa the short-styled flowers have a larger and 

 more striking limb to their corollas than the other forms in each case. It may 1 •■ 

 taken to be a general rule that the flowers adapted to cross-fertilization, in which 

 no autogamy takes place, are larger than those in which the accomplishment of 

 autogamj' is assured. This phenomenon has been explained by the circumstance 

 that flowers destined to be crossed with others require to be more plentifully 

 equipped with the means of attracting insects than those which are certain to 

 undergo fertilization even if no insects visit them. 



The means whereby autogamy is achieved in species with heterostyled Sowers 

 are in the main the same as they are where the flowers are not heterostyled. In 

 some cases the stamens, or the petals with pollen affixed to their surfaces, elongate 

 sufficiently to enable the stigmas to come into contact with them: in others the same 

 result is attained by an inclination or an inflection of filaments or style: in a third 

 series the stigma is dragged through the ring of anthers when the corolla falls off, 

 or the petals by opening and closing effect the transference of the pollen from the 

 anthers to the stigma; and lastly, in some instances, the pedicels undergo elongatioD 



