the Determination of Sex in Plants. 431 



is provided for in the reduction division 1 , we may consider certain 

 corollaries which might be expected to follow from Castle's 

 hypothesis. 



In plants the problem of sex-determination is rendered more 

 complex by the existence of two separate generations, the gameto- 

 phyte and sporophyte, which alternate with one another in the life 

 history of the plant. 



In the higher Cryptogams and in Phanerogams the gameto- 

 phyte is dioecious, whatever state prevails as regards the distribu- 

 tion of megasporangia and microsporangia upon the same or 

 different individuals of the asexual generation ; but in many of 

 the Archegoniatae the gametophyte is hermaphrodite, bearing 

 both antheridia and archegonia upon the same thallus. 



Campbell ('95 ; p. 335) was not certain whether a reduction 

 in the number of chromosomes takes place at the formation 

 of the spores in Ferns ; and while Strasburger ('94) previously 

 gave the number of chromosomes in Osmunda as 24 and 12, in 

 the sporophyte and gametophyte respectively, Guignard ('99) has 

 observed that sometimes at least 22 chromosomes were to be 

 counted in the spores. Strasburger ('00, a; p. 77) confirms this, 

 at the same time stating that the number, 12, which he had given 

 as that present in the gametophyte, was correct for those 

 examined. Further, while Calkins ('97) described some Lepto- 

 sporangiate Ferns as possessing about 128 and 64 chromosomes, 

 Stevens ('98) gave the numbers as 64 and 32. 



A possible extension of Castle's hypothesis in this connexion 

 was suggested to me by Mr Bateson. It might be supposed that 

 no chromatin reduction takes place at the formation of spores 

 which are destined to give rise to hermaphrodite prothallia ; that 

 it only occurs at some later mitosis previous to the formation of 

 the gametes. That is, the nuclei of unisexual prothallia would 

 possess the reduced number of chromosomes ; those of prothallia 

 actually or potentially hermaphrodite would contain the same 

 number as occur in the nuclei of the sporophyte*. It seemed 

 possible that in this hypothesis lay the explanation of the con- 



1 With, however, the modification that the chromosomes corresponding to the 

 dominant and recessive of each pair of alleloniorphic characters will he separated 

 in such a way that any gamete may contain any combination of characters, so long 

 as allelomorphs are separated, and not, as Cannon allows it to be supposed, so that 

 all the chromosomes derived from the father pass to one daughter nucleus, and all 

 those from the mother to the other daughter nucleus. As Sutton has pointed out 

 the latter hypothesis does not provide for the observed combinations which occur 

 in the offspring of hybrids between two races differing from one another in respect 

 of more than one character. 



2 Farmer ('94) had found reduction of the chromosomes at spore formation in 

 certain Hepaticae. The want of uniformity in the nomenclature of this group left 

 room for doubt as to whether the gametophyte was hermaphrodite in any of the 

 species described by that author. 



