OBSERVATIONS UPON MULING AMONG PLANTS. 



165 



deltoides J , and some others, there is a little difference in size, 

 but whether absolutely dependent on impregnation does not 

 appear very clearly. 



Attempts to hybridize species of CEnothera with rough and 

 angular seeds, with those possessing smooth and round seeds, 

 did not succeed ; there, however, seems to be little disposition 

 amongst many species to hybridize. It has been supposed that 

 hybrid impregnation has an effect upon the colour of the seeds 

 of peas and other leguminous plants, and differences in colour 

 certainly occur in peas which seem to be due to that cause, 

 even taking into consideration the facility with which variations 

 arise in plants of which so many mere forms exist. The case, 

 however, is different as regards varieties of Indian corn. The 

 white dwarf variety, when inoculated with the large red-striped 

 form, produces seeds which exhibit no change of colour, though 

 when sown they give rise to variegated spikes. The pea at pre- 

 sent is almost the only exception to the law that the influence of 

 strange pollen produces no immediate change in the peculiar 

 form and outward properties of the fruit and seeds, but merely 

 in the capacity of the embryo to produce, when developed, 

 modifications of the respective species, and in the form of their 

 parts. 



The altered quality of the embryo does not appear outwardly, 

 nor can it be discerned by the microscope. That this modifying 

 power exists in the pollen does not admit of doubt, and the con- 

 stancy of hybrid types produced between pure species shows that 

 its tendency is in a great degree definite. It has, indeed, been 

 supposed that variations and half-hybrids are produced in pro- 

 portion to the degree of admixture of strange and native pollen, 

 and that the paternal or maternal type prevails according to the 

 prevalence of the paternal or maternal element ; but the author's 

 experiments do not encourage such notions, though they seem to 

 rest upon good authority.* A certain quantity of pollen is neces- 

 sary for perfect fecundation ; and in cases of mixed fecundation the 

 types are those of the pure matrix or of the normal hybrid. In 

 some cases, too, of hybrid fecundation, the maternal type always 

 predominates ; in others the paternal ; sometimes to such an 

 extent, that the deviation of the typical hybrid from the pure 

 parent on either side, as the case may be, is very slight. 



The principal facts respecting imperfect fecundation have 

 already been mentioned incidentally ; we pass on, therefore, to 



* Such is the theory of Mr. Thwaites in his paper on the conjugation of 

 Diatomacece quoted above, which would be borne out were it certain that he 

 had to deal with pure species. It is probably at the present day almost 

 impossible to get pure individuals except of very recently introduced 

 species of Fuchsia. 



