SEMI-CLONAL REPRODUCTION 463 



telophase by the seven daughter bivalents going to this pole. In 

 the pentaploid forms there are therefore usually 28 chromosomes 

 at the micropylar end and 7 at the chalazal end. The second 

 division is regular and yields two small nuclei and two large ones ; 

 one of the latter gives the functional embryo-sac, as is shown by the 

 chromosome numbers of hybrid seedlings. 



Self-fertilisation of these plants will therefore yield offspring like 

 the parent having 7 pairs of chromosomes of identical origin and 

 21 chromosomes that are not directly related to them or to one 

 another. A complement of a race of such a kind consists of 7 pairs 

 of chromosomes that are transmitted by sexual reproduction (and 

 no doubt inbred and homozygous) together with 21 chromosomes 

 that are transmitted virtually by vegetative processes, for they are 

 neither separated by meiosis nor brought together by fertilisation. 

 Reproduction, in fact, may be described as semi-clonal. In regard 

 to its genetic effect this mechanism is analogous with poly- 

 ploidy and with the ring formation of complex heterozygotes, 

 inasmuch as it enables a hybrid to breed true. The mechanism is 

 responsible for the relative constancy of the innumerable species of 

 Caninae. It is also responsible for the ability of widely distinct 

 forms to hybridise. The seven chromosomes of the pairs are 

 probably highly constant throughout the group. Their pairing 

 alone is necessary for fertility. The unpaired chromosomes are, 

 on the other hand, the organs of variation. In this specialisation of 

 function the two groups of chromosomes are analogous to the 

 differential and the pairing segments of the chromosomes of the 

 heterozygous CEnothera species. 



The origin of the system does not present much difficulty. 

 Repeated crosses amongst diploid species or a cross between two 

 hexaploid species would give the condition with seven pairs of 

 chromosomes and twenty-eight univalents, as in Triticum-Mgilops 

 crosses. It is not now necessary to suppose, as Tackholm did, a 

 cross between a diploid and an imaginary decaploid as the source of 

 such a form. The hybrid would give at once a proportion of pollen 

 grains with only the seven chromosomes derived from the pairs. 

 The essential new property of these forms is therefore the 

 polarisation of the embryo-sac mother-cell whereby it regularly 



