May 1, 1895.] 



KNOWLEDGE 



103 



guests are prevented from reaching tbe nectar by the 

 depth of the corolla-tube ; its small diameter, too, makes it 

 impossible for any insect to rifle the flower without touching 

 the stigma and anthers. Contained in the base of the 

 corolla-tube, the nectar is reserved for bees and butterflies 

 having a slender proboscis long enough to reach to the 

 bottom of the flower. A visitor alighting on a long-styled 

 flower thrusts its proboscis down to the bottom of the 

 corolla-tube ; the middle part of the proboscis comes in 

 contact with the anthers placed in the middle of the tube, 

 and is coated with pollen. Should the insect next visit_ a 

 short-styled flower, the pollen-dusted part of the proboscis, 

 when inserted into the tube, is level with the stigma, and 

 some of the pollen is almost certain to adhere to its viscid 

 surface. At the same time, the basal part of the 

 proboscis is dusted against the anthers situated at the 

 top of the tube, and this pollen will be transferred 

 to the stigma of the next long-styled flower which 

 the insect happens to visit. By conveying pollen on 

 different parts of its proboscis, an insect will thus 

 constantly fertilize the long-styled flowers with pollen 

 from the short-styled ones, and i-ice rers,!. The crossing 

 of the two forms in the manner described, which is 

 the most natural result of insect visitation, Darwin 

 distinguished as "legitimate"; the fertilization of a flower 

 of either form with pollen from a flower of the same form, 

 he termed " illegitimate crossing." He found, on artificially 

 fertilizing the flowers, that the legitimate method was 

 much more productive than the illegitimate — the average 

 numbers of seeds being in the proportion of 100 to 54. 

 Seeds resulting from illegitimate unions were, moreover, 

 smaller than the others, and gave rise on germination to 

 inferior seedlings. Convincing proof was thus firnished 

 that the long and short-styled flowers of Primula are 

 adapted for reciprocal cross-fertilization. This result has 

 been confirmed by several investigators, but it is well to 

 remember that it is deduced from averages, and that these, 

 when analyzed, ofler certain anomalies for which it is not 

 easy to account in the present state of our knowledge. 

 The figures obtained by experiments on different plants 

 vary a good deal ; nevertheless, the broad fact is established 

 beyond question that the maximum fertility, as shown by 

 the number and size of the seeds, is only obtained when 

 dissimilar flowers are intercrossed in the waj' this would 

 naturally be accomplished by an insect visiting the 

 flowers. 



The common primrose, cowslip, oxlip, auricula, poly- 

 anthus and many other primulas, including the common 

 table plants P. sinensis and F. oJiconicn, are all dimorphic. 

 Our illustration is taken from the last-mentioned species. 

 The water-violet [Hottonia palustris) is another of the 

 Primulacese with dimorphic flowers ; these were arti- 

 ficially fertilized by Herman Miiller, and his results agree 

 with those obtained in the case of Primula. Taking the 

 result of legitimate fertilization as 100, the efficiency of the 

 illegitimate mode would be represented by 65-5 ; of inter- 

 crossing flowers belonging to the same individual plant by 

 5'6 ; and of self-fertilization, or the impregnation of flowers 

 with pollen from their own stamens, by 11-4 per cent. 

 Some of the gentian order also bear dimorphic flowers. 

 The beautiful white spires of blossom with which, in May, 

 the bog-bean {^lenijnntlws trifoliata) enlivens our marshes, 

 consist of long and short-styled flowers. Tbe curious fact 

 is recorded by Warming, that in Greenland this plant has 

 become homostyled. Another of the gentian family, the 

 centaury (Erythrrm centnurium) blooming towards the end 

 of summer, is, as I have observed, at least occasionally 

 dimorphic. The purple-flowered flax [Linum grandi/Iomm), 

 a common garden annual which must be known to most of 



our readers, is heterostyled ; its flowers are remarkable as 

 being absolutely barren when fertilized in tbe illegitimate 

 manner. The lungwort {Puhnunaria ojticiiialis), one of the 

 borage order, and the buckwheat {P(jli/!ionum fafiopyrum) 

 may also be mentioned as examples of dimorphism. The 

 order Kubiaceas, of which the woodruff may be taken as a 



Dimoi'pliic Floirovs of Primrose (^Primula otimiiira.) A, Long- 

 stvled form, with small pollen and large stiginatic papill;i?. a, 

 anthers ; s, stigma. B, Sbort-stjled form, with large pollen 

 grains and small papilhc. n, insect's prohoscis dusted in B, m, 

 ditto in A. Pollen and papilhc highly magnified. 



type, is, however, the richest in dimorphic species ; these 

 are all foreign, and several of them, though dimorphic in 

 form, really have the sexes completely separate, the short- 

 styled flowers having become male, and the long-styled 

 ones female. 



Great as is the interest attaching to the dimorphic class, 

 certain other plants, which produce three different sorts of 

 flower, excel them. The purple loosestrife (Li/t/intm 

 salicaria), which grows in this country both wild and 

 cultivated, is a typical example of this trimorphism. Its 

 dissimilar flowers are borne, as in the dimorphic class, on 

 separate plants, and are distinguished as short-styled, mid- 

 styled, and long-styled. Each variety has two sets of 

 stamens differing in height, the anthers being placed at 

 levels corresponding to the stigmas of the other two 

 forms ; the short-styled flower has medium and long stamens, 

 the mid-styled long and short stamens, and the long- 

 styled short and medium stamens, as shown in the 

 accompanying figures. As the long and medium stamens 

 of the loosestrife project beyond the flower, they are more 

 likely to come in contact with the body than with the 

 proboscis of a visitor. The bee Cilissa melanura shows a 

 singular partiality for the loosestrife, and confines its visits 

 almost exclusively to flowers of this species. The 

 dimensions of the insect accurately correspond with those 

 of the flower, and H. Miiller has noted that when engaged 

 in extractuig the nectar, it touches the shortest organs with 

 its head, the intermediate ones with the ventral surface of 

 its thorax, and the longest with the ventral surface of its 

 abdomen. Darwin's numerous experiments on this plant 

 conclusively show that for complete fertility it is essential 

 that a stigma should receive pollen from a stamen of equal 

 height. Of the eighteen possible modes of fertilization 

 only six were found to be fully productive — viz., when the 

 short-styled form was pollinated from the short stamens of 

 the mid or long-styled flowers ; when the mid-styled was 

 polHnated from the medium stamens of the long or short- 

 styled blossoms, and when the long-styled was pollinated 

 from the long stamens of the short or mid-styled flowers. 



Upwards of twenty species of Lythrum are known to be 

 heterostyled, but a much larger number have flowers of the 

 ordinary form. The genus Oxalis also contains numerous 

 trimorphic species. The wood sorrel of our plantations, 

 with whose pretty heart-shaped leaflets and pinkish white 

 flowers our readers must be familiar, is homomorphic, but 

 several closely allied exotic species bear three kinds of 

 flower. One species of Oxalis is frequently met as an 

 accidental introduction in greenhouses ; the lengths of the 

 styles and stamens in its different flowers are delicately 



