104 



KNOWLEDGE. 



[May 1, 1896. 



adjusted with a degree of accuracy which might well be 

 described as mathematical. Pontederia, an American 

 aquatic, is remarkable as afifi'ding an example of a 

 trimorphic monocotyledon. Epigsea, one of the heath 



Trimorpliic Flowers of Loosestrife {Li/lhntm salicaria). a short, 

 b mecluiiy, c long stamens; ,t, stigraa. 



order, attains, however, the highest degree of hetero- 

 morphism known. This plant, according to Asa Gray's 

 description, is tetramorphic, producing four different sorts 

 of flower, with styles and stamens of four different 

 dimensions. 



Heterostyled flowers differ not only as regards the 

 length of their stamens and pistils, but in other respects 

 as well. Their corollas sometimps differ considtrablv in 

 size ; short-styled flowers commonly produce larger pollen 

 grains, and the stigma of the long- styled flower is larger 

 and beset with longer papillae than that of the short-styled 

 blossom. From these differences it has been inferred 

 that dimorphic species are on the way towards becoming 

 dicEcious ; in other words, are tending towards a complete 

 separation of the sexes, a condition exemplified in the 

 nettle, willow, poplar and many other plant?. Long-styled 

 plants have been shown to be more prolific than the others, 

 whether the fertilization be legitimate or illegitimate ; the 

 intercrossing of two short-styled forms is, on the other 

 hand, characterized by remarkable and, in some cases, 

 absolute sterility. This circumstance has been held to 

 indicate that the pistil of the short-styled form has tmdcr- 

 gone partial degeneration, but a strong objection to this 

 view is the fact that the short-styled flower generally excels 

 the other in its capacity for self-fertilization. la several 

 experiments the short-styled flowers wtre found to yield 

 abundance of seed when fertilized with their own pollen. 

 The mid-st}led form of Lythrum is the most fertile both in 

 legitimate and illegitimate unions, but it is entirely unpro- 

 ductive with pollen from another midstj led flower. The 

 longest stamens of loosestrife have the largest pollen 

 grains, their colour being green ; the pollen of the short 

 and medium stamens is yellow. Its stigmatic papillae 

 difler in a manner similar to those of dimorphic species. 



The results of artificial fertilizition fluctuate very much ; 

 Hildebrand, for example, found the lungwort completely 

 sterile with its own pollen and with that from a flower of 

 the same form. Tne long-styled variety of this plant 

 Darwin, on the other hand, found highly productive, even 

 when illegitimately fertilized, and he also obtained a number 

 of seeds as the result of self-fertilization. If this dis- 

 crepancy arose from Hildebrand's plauts being kept indoors 

 while Darwin's were grown outside, it would appear to 

 show that the capacity of forming fertile unious is a 

 character subject to great variation and liable to be afi'ected 

 by alteratious in the external conditions under which 

 plants are grown. 



A noteworthy fact, revealed by investigations on this 

 subject, is that flowers of the same species when illegiti- 

 mately united behave in exactly the same manner as do two 

 distinct species when crossed. If pollen from a different 



species be applied to a flower's stigma, and its own pollen 

 be afterwards placed on the same stigma, the latter is so 

 strongly prepotent that it generally arrests the efifects of 

 the foreign pollen. In heterostyled flowers legitimate pollen 

 is also strongly prepotent over illegitimate when placed on 

 the same stigma. The effect of illegitimate fertilization 

 was found to be completely neutralized by the application, 

 twenty-four hours later, of legitimate pollen. Again, the 

 ofl'spring of illegitimate unions admit of being crossed in a 

 legitimate manner, but when this is done, instead of the 

 complete fertility which we should expect, such unions 

 exhibit marked and, in some cases, absolute sterility. 



There is thus the closest analogy between these illegiti- 

 mate plants and hybrids. As regards fertile union, a 

 short-styled primrose stands related to another flower of 

 the same form, very much as it does to a flower of a 

 difl'erent species altogether, and an illegitimate plant 

 derived from the crossing of two similar primroses is, as 

 regards reproduction, to all intents and purposes a hybrid. 

 The sterility of hybrids was long regarded as an unfailing 

 test of species ; the phenomena of dimorphism and 

 trimorphism, however, foibid us to regard this as a safe 

 criteiion of specific distinction. 



The varying degrees of fertility disclosed by investigations 

 of this kind, lend countenance to the theory advanced a 

 few years ago by the late G. J. Eomanes to explain the 

 origin of species. It is not every variation, as he pointed 

 out. that can give rise to a new species, for natural selection 

 only accounts for the preservation of useful characters. 

 Allied species are commonly distinguished by characters 

 which are of no physiological importance whatever. 

 Further, when a new variation has once arisen, there is 

 nothing in natural selection to prevent its disappearing 

 again through intercrossing with the parent stock. To 

 obviate these difficulties, Eomanes brought forward his 

 supplementary theory of physiological selection, or the 

 segregation of the fittest, which attributes the origin of a 

 new species to the appearance of a variety which is sterile 

 with the parent form, although fertile when intercrossed 

 with others of its own variety. Such sterility, when it 

 arose, would prove a barrier against retrogression. 



Since their near allies are all homomorphic, the ancestral 

 forms of heterostyled plants doubtless bore but one kind 

 of flower. The st}les and stamens of many flowers are 

 suljtct, especially under cultivation, tj vary much in 

 length. Selection by insects would, however, tend to fix 

 the stamens and styles at the proper height, and the 

 flowers might thus ultimately become dimorphic or 

 trimorphic. As heterostylism occurs in families which 

 have but little affiuity with each other, the arrangement 

 has in all probability arisen independently in different 

 groups. 



Cross-fertilization is in many flowers promoted by 

 dichoijamy ; the stamens and stigmas do not mature 

 simultaneously, but the flav;ers are alternately male and 

 female. Where this arrangement exists, two insect-visits 

 are required for the proper fertilization of each flower. 

 Over these, heterostyled pliuts have this advantage, that 

 one insect-visit suffices both for the pollination of the 

 stigma and the removal of the pollen. Where the proper 

 insects are scarce, this advantage may become of some 

 importauee. It is right to mention, in conclusion, that 

 Heuslow takes exception to some of Darwin's deductions 

 regarding the advantages of cross-fertilization, and points 

 out that uoder certain circumstances, self-fertilization 

 appears to be quite as effectual. Fresh experiments would 

 seem to be required to ascertain precisely the conditions 

 under which each method is adopted. Henslow's views 

 I are to a certain extent, however, in accord with Darwin's ; 



